The report, written with Advance HE, draws on 244 staff and 91 students and student representatives from at least 127 institutions. They took part in roundtables, workshops, interviews and online prompts between June and September 2026. The report says the participants chose to take part, so they may not represent all students and staff.

The report says many unsupervised assessments can be completed to a high standard with AI tools. Students can do this without showing the knowledge the assessment is meant to test. Some students told researchers that peers finished assessments with AI in minutes and got similar grades to those who did the work. The report also says AI detectors do not reliably identify AI use.

Alongside the report, the OfS listed approaches it considers likely to comply with its quality conditions. These include secure assessments designed to prevent AI use, such as in-person exams, presentations, vivas and practical tasks. Courses should also keep up with AI in their subject. For computer science, its example is considering what machine learning means for computing. Times Higher Education reported that the regulator stopped short of requiring AI-free assessment.

For students, the OfS expects each institution to say clearly whether and how AI may be used in each assessment. The report found many students struggled to understand their university's AI rules. Students also asked for more AI skills training, which the report says is uneven. The OfS says it can fine institutions that fall short or suspend their funding.

The OfS rules apply only in England. The report gives no figure for how many assessments can be completed with AI.

The Charter of Engagement for Safer AI for Children, also published by the French presidency, says it has been endorsed by Anthropic, Black Forest Labs, Microsoft and OpenAI. It sets up a research network led by an independent scientific committee. Methods and results will be published, including null or negative findings. Companies may audit how their systems are tested, but the charter says this gives them no right to delay or suppress findings.

The Paris Peace Forum said that AI developers such as OpenAI and Anthropic would join the nonprofit Common Sense Media in pledging €8m for research on AI and child development. The charter also asks for age checks that do not rely only on users stating their age. It calls for age-appropriate AI literacy for children.

The same day, more than 50 countries signed a separate Children First Declaration, the forum said. They include the UK, Australia, Canada and South Korea, along with UNICEF and UNESCO. The declaration covers five areas to 2045, among them education and children's wellbeing in the digital age.

Before the forum, AFP reported that a French expert committee was due to give President Emmanuel Macron advice on AI's effects on children's mental health, development and education. AFP said the charter is meant to inform national and European law-making.

The charter and the declaration are voluntary pledges, not laws. They set no new rules for schools. The charter says long-term research on children's use of generative AI is not yet available. For parents and teachers, its main promise is independent, published evidence on how AI affects learning and development.

The forum has not said how the €8m is shared between the funders. It is not known when the scientific committee will be named or when findings will appear.

The American Federation of Teachers and the United Federation of Teachers said in their announcement that the standard applies at once to Anthropic's existing contracts with school districts that use Claude for Teachers. That product, launched in July, gives verified US school teachers free access to Claude. The unions said student and teacher data may be used only for authorised educational purposes. It can never be sold or used for targeted advertising.

The agreement also says teachers, not AI systems, stay responsible for decisions about students. Schools and families must get plain information about what each product collects and where it falls short.

Education Week reported that the terms match a standard Microsoft signed in September, which took effect on 1 November. It said the Anthropic deal is narrower because Claude for Teachers has no tools for pupils. It also does not cover other companies' products built on Claude. K-12 Dive reported that districts can add the Microsoft terms to their contracts without renegotiating them.

For teachers in the United States who use Claude for Teachers, the protections now apply to their district's existing contract. Education Week said districts will still need to opt in for future contracts. The agreement does not cover schools in the UK.

Elizabeth Kelly, Anthropic's head of beneficial deployments, said the company would build on the agreements "for our education approach globally". She did not say how.

Education Week said the unions are still negotiating with OpenAI. It said the AFT has no agreement with Google or Meta. The unions' announcement does not say how compliance will be checked.

Teacher Tapp runs a daily survey app answered by more than 10,000 teachers. In its weekly results, it asked teachers who had set homework needing a device or internet in the past month whether any pupil failed to do it for lack of access. Half of secondary teachers and 37% of primary teachers said at least one pupil had.

In secondary schools with the most pupils on free school meals, 31% of teachers said this had happened to several pupils. In the most affluent schools the figure was 7%. The primary figures were 19% and 6%.

The Digital Poverty Alliance, a charity, found a similar problem in July. In its census, 30% of teachers said at least 10 of their pupils lacked a suitable device or reliable internet at home. Fewer than one in 10 said pupils could do all their coursework on a smartphone.

Teacher Tapp also asked about screens in lessons. Most secondary teachers said pupils used no computer or tablet in their first lesson of the day. Most primary teachers said their pupils had no screen time at all that day. BBC News reported on 4 November that the government is drawing up screen time guidelines for schools and parents.

Online homework is now common. Only 12% of secondary teachers and 45% of primary teachers said they had set none in the past month. For pupils without a laptop or internet at home, that work is harder to complete. The figures show the problem is most common in deprived areas.

Teacher Tapp's panel is not a random sample. The post did not say how many teachers answered each question. It is not known how many pupils are affected in total.

Schools already follow statutory guidance that asks them to be phone-free by default. Many meet it with a "not seen, not heard" rule, where a phone can stay in a pupil's bag. The department said in its announcement that stronger guidance will ask schools to go further. It listed lockable pouches, collecting phones at the start of the day, secure storage and banning phones from the site.

BBC News reported that the stricter guidance will apply from September 2027. The department said only 16% of secondary pupils say phone rules are followed all of the time. Its own survey from June found 23% of secondary parents who knew the policy said their child must hand in or lock away a phone, up from 9% a year earlier.

School leaders' unions warned about cost. Paul Whiteman of the NAHT told Tes that schools would need extra funding and support. Pepe Di'Iasio of the Association of School and College Leaders said storage for every pupil was "not feasible" without government money. Tes said the announcement did not mention new funding.

For pupils and parents in England, a phone kept switched off in a bag will no longer be enough once the new guidance applies. Each school will still choose its own method, so the rules will differ from school to school. The change covers England only. The BBC said Scotland, Wales and Northern Ireland are each consulting on their own approach.

The department has not published the new wording of the guidance. It has not said whether schools will get money for pouches or lockers.

The consultation document names four kinds of product: web filtering, products that do high-risk processing of children's data, generative AI tools that pupils use and personalised learning tools. It says AI tutors, SEND platforms and school management systems are likely to count as high-risk data processing. Office software, PDF textbooks, hardware and AI tools used only by staff would be left out.

AI tools for pupils would be checked against eight of the department's existing generative AI safety standards. These include limiting the risk that pupils lose skills by relying on AI, avoiding emotional dependence and not using manipulative design. The checks would cover safety only, not whether a product helps pupils learn.

Accredited independent bodies would carry out the checks. The department proposes full checks every three years. Generative AI tools could need them yearly. It says setting up the scheme would take at least 18 months before any product could be certified.

Tes reported that the plans follow a minister's warning about a "Wild West" of edtech products. In the Commons on 10 November, education minister Josh MacAlister said AI tutors should not be widely used before their "full effects on cognitive development" are known.

For schools, nothing changes yet. The department says it has no power to make schools use certified products. It proposes an expectation in statutory guidance instead.

If the scheme goes ahead, the department would publish a register of certified products, for example on GOV.UK. AI tools that a teacher builds for use only in their own school would not need a certificate. Teachers, parents and pupils can respond until 4 February 2027.

The department has not said what certification would cost suppliers or named a body to run the scheme. It plans to publish its response in spring 2027.

The Teachers' Union of Ireland (TUI) and the Irish Federation of University Teachers (IFUT) held a symposium on AI and assessment that day at the Grangegorman campus of Technological University Dublin. In a joint statement, they said that relying only on detecting AI or banning it is now outdated.

The unions listed five demands. They want a say in decisions about AI and a review of how assessment is designed and run. They also want large investment in staff training and clear guidelines that are revised regularly.

Their fifth demand is one national approach, so that no student is put at a disadvantage.

The statement said continuous assessment is under particular pressure. It has long been central to Irish higher education. Michael Gillespie, general secretary of the TUI, said many members are worried about the integrity of assessment.

RTÉ reported that lecturers at the meeting described students using AI as a shortcut for assignments and essays. Some said it harms critical thinking. Others said it can be used creatively. Both unions agreed that AI should not be banned, RTÉ said.

For students and lecturers in Ireland, no rule has changed. The statement sets out what the unions will ask of policymakers and colleges. They did not say which body should lead a national approach or what new forms of assessment should look like. Neither the statement nor the RTÉ report included a response from the government or from the colleges.

All the sources for this story are in German. The translations are our own. The ministry calls the strategy a framework for orientation. The Austria Press Agency reported that it sets no legal rules or bans and no list of required steps. The agency said decisions on how AI is used in lessons stay with teachers.

Wiederkehr said pupils should first learn without AI. Next they should learn about AI and how it works. Only then should they learn with it. The ministry's summary says basic skills are built and shown without AI.

According to the summary, teachers may use AI to plan lessons, adapt tasks and give feedback. Its output counts only as a draft. The summary cites research that nine in 10 pupils in Austria use AI chatbots. Almost three in four of them use them for school work.

The ministry lists support that will follow. Teachers can already use a FAQ and guidance. An advice service on AI for schools and a set of checked AI tools for teachers are due from the summer semester of 2027. An AI sandbox, a virtual space for learning about AI by experimenting with it, will be tested from early 2027.

Reactions were mixed, the press agency reported. Heinrich Himmer, education spokesman for the Social Democrats, called it a right step but wants binding rules. The Austrian Economic Chamber said it lacks timelines, funding and targets that can be checked.

For teachers and parents in Austria, the strategy changes no rules now. In September, Wiederkehr told the press agency that AI means the rules on grading should change, so that schools can assess pupils more flexibly. He said the governing coalition must discuss this first. The report gave no date. The strategy document will be updated in regular cycles.

The plan is part of a new Education and Skills Package. The Commission said in its press release that it will first write learning outcomes and assessment guidance for AI literacy, suited to each age. They will be based on the AI Literacy Framework that the Commission and the OECD published in June 2026. The Commission's policy paper says the learning outcomes are due by mid-2027. It also promises EU funding for long-term research on how AI affects children's thinking and results at school.

The package has other measures on digital learning. On 3 November the Commission set up EDU-EDIC, a group of eight member states that will share trusted AI and digital services for schools. The Commission also plans a Healthy Screens Pledge. It would give schools and families guidance on online safety and screen time.

Euronews reported that the certificate is meant for 15-year-olds. The Commission's press release, questions and answers and policy paper give no age. Euronews quoted Roxana Mînzatu, the executive vice-president who presented the package, as saying the certificate would help pupils understand how AI works and spot manipulation.

For schools, nothing changes yet. The Commission says in its questions and answers that the package creates no binding obligations. National governments will decide whether to adopt two non-binding recommendations, one on basic skills and one on teachers. The second asks governments to give teachers a say in how AI is used.

The Commission has not said which countries will test the certificate, how pupils will be assessed or whether taking it will be voluntary. It will discuss the package with governments and teachers' organisations at a summit in Brussels on 20 November.

Google said on 3 November that schools that ban its AI will lose Search on pupil accounts. After criticism that day, it said on X that schools can turn on Search alone, without Gemini. NBC News reported that Search will still include AI features, which Google said will guide pupils through problems step by step.

AFP reported that New York City's school system has already disabled Gemini and AI Mode on pupils' Chromebooks. It has asked Google to disable AI Overviews, the summaries above search results. Los Angeles Unified said it is reviewing the announcement.

Mayor Zohran Mamdani accused Google on X of using "its monopoly power" to force AI into schools. Senator Bernie Sanders called the plan "extortion", TechCrunch reported. Education Week said Google's statement did not answer nine questions from two US senators.

For school leaders, the choice is Search with AI or no Google Search on pupil accounts. Education Week reported that settings can differ by grade and that outside tools can block some AI content. It quoted Common Sense Media, a children's advocacy group, as saying smaller districts may lack the staff for this. Rich Calkins, a superintendent in New York state, told K-12 Dive his district is waiting to see how Google applies the change.

Google told AFP the controls will be rolled out worldwide, beginning in the United States. It has set no date, though Education Week reported that Google expects the changes early next year. In Australia, ABC News quoted a Swinburne University AI expert objecting to the Search condition.

Google has not said when the controls will reach UK schools. Education Week said its statement did not say whether Search without AI is coming. Chalkbeat reported that New York officials did not answer whether the city would consider leaving Google products.

The communique from their meeting on 2 November says ministers will also "introduce, trial or consider" limits on screen use in schools. States, territories and school sectors will decide how to apply these measures. The communique sets no national limit on screen hours. It does not say which uses of AI count as restricted.

A new Australian Schools Digital Detox Taskforce, led by an independent expert, will advise ministers. Jason Clare, the federal education minister, said in a speech on 5 November that it will survey students, teachers and parents and advise on further action next year. He said there is "a bit of evidence" that AI can undermine children's thinking skills.

Reports differ on how far the limits go. CathNews, summarising The Australian, said primary pupils will be blocked from generative AI in class. The communique uses the word "restrict" and leaves the detail to each state. ABC News quoted Clare describing a digital detox that restricts screen time and AI.

National tests will change too. The Year 5 writing test in NAPLAN, Australia's national assessment, will return to handwriting in 2028, subject to advice on feasibility and cost. In September, more than 85 experts and organisations asked governments to look at how screens are used in school, not only for how long.

For teachers and parents in Australia, the practical rules will come from each state or school sector. Victoria has already said that pupils in Years 3 to 6 will spend at most 90 minutes a day on screens from 2027. The communique does not name the taskforce's leader. It also does not say when each state will act.

EU Code Week is a campaign, supported by the European Commission, that gets young people to try coding. Its announcement said the round had €19,836 to give. Each project could receive up to €9,918. It received 553 applications from 24 countries, including 347 from Türkiye and 101 from Ukraine.

NGO uBoost in Ukraine won €9,918 for a project that mixes coding and AI lessons with advice on careers in science and technology. It will run school activities, bootcamps and online sessions. It focuses on girls, young women and learners who face barriers.

The Education Technology and Game Association (ETODER) in İzmir, Türkiye, won €8,706.50. Its project will run coding, robotics and AI activities for students, teachers and families. It will also train teachers and make resources they can reuse.

JA Europe ran the call for the Code4Europe group of partners. The money comes from the Digital Europe Programme, which the European Commission describes as an EU funding programme with a budget of more than €8.1 billion. The Commission lists advanced digital skills among its aims.

The two projects run from 29 October to 1 December 2026. The announcement does not say how many young people the projects expect to reach. It also gives no dates for another round of grants.

On the CSUK One home page, CSUK said the platform brings its courses, coding tools, quizzes and assessments into one place. It listed a Code Studio for Python, HTML and block-based coding. The studio has Parsons problems, in which a learner puts given lines of code in the right order. It also has debugging challenges, optional AI marking and an AI tutor.

CSUK said anyone can start with a free account and that no paid membership is required. Teachers can create classes and use free daily activities, it said. Its shop page lists a "Complete School Bundle" covering KS3, GCSE, coding and teaching resources for £650.

For teachers who already pay for CSUK, the company said CSUK One will recognise active memberships. It said they unlock the matching areas for the rest of the membership, so schools do not buy them again. CSUK said a transfer process will move classes without setting up each pupil again. Past pupil results and progress will not transfer because the courses were rebuilt. The CSUK:Teacher site says teaching resources have moved to CSUK One and that the site closes at the end of term.

CSUK did not give an exact closing date for CSUK:Teacher. Its pages also do not say which exam board specifications CSUK One covers or list what the £650 bundle includes.

Bruno Reddy, founder of the company behind it, told BBC News on 4 November that every school will have a curfew and a daily limit turned on by default. Leaderboards will be off by default. The company says the app and a sister program, Numbots, are used in 17,000 schools.

The change follows a hearing of the Commons Science and Technology Committee on 3 November. The committee is looking at how screens affect children's brain development. Emily Darlington, Labour MP for Milton Keynes Central, said the app rewards pupils for how long they play rather than how well. Pupils earn virtual coins and rock star outfits.

Paul Waugh, the minister for early education, answered her at the hearing. He said the claim that gamification helps learning is "propaganda" from edtech firms and that it is about driving screen use. Tes reported that he warned of a "Wild West" of edtech products for schools. He said the government is testing products with teachers.

Reddy told the BBC he shared concerns about screen time. He said the company advises schools on about three minutes a session, four or five times a week. A typical child played on three days in September, for about six minutes each time, he said.

For schools that use the app, the new defaults start on 24 November. The TT Rockstars website says a 20-minute daily limit is already on by default for play outside school hours. Schools can also set a curfew, a time when pupils cannot use the app. The BBC reported that the government is writing screen time guidelines for schools and parents.

It is not clear whether the new school default will use the same 20-minute limit. The government has not said when its screen time guidelines will be published.

Quincy Larson described the tool in a post on the freeCodeCamp news site. He said Socrates first looks through the thousands of hints and test cases that freeCodeCamp staff have written. A test case is a small check that shows whether a learner's code does what the challenge asks. If none of those fit, Socrates writes a new hint.

freeCodeCamp's code page for Socrates on GitHub says the tool receives the learner's code, the challenge description and the tests that are failing. It then returns a hint that points towards the fix without giving the solution. The page lists separate prompts for HTML, CSS, JavaScript and Python. Its default AI model is openai/gpt-oss-20b.

For learners, freeCodeCamp said Socrates is live now and can be used inside the lessons at freecodecamp.org/learn. The post also said the freeCodeCamp forum is now focused on discussion rather than questions and answers. freeCodeCamp said it expects most learners to get unstuck with Socrates. It wants the forum used for talk about hobbies, studies and goals.

The post did not say whether there is a limit on how many hints a learner can ask for. It also did not say how freeCodeCamp checks the hints that Socrates writes for mistakes.

She made the comments at her first hearing before the House of Commons education committee. FE Week reported that the government wants new and existing "practical" GCSEs, including computing and drama, to include "a lot more practical assessment". It said the Department for Education will also design new technical qualifications around practical assessment.

Tes reported that Powell named drama, music and computing as subjects that need more practical assessment to "make them fit for purpose". Asked how AI could be used in assessment, she raised concerns about its use in marking. She said marking should be done by "the human eye" because AI could "exacerbate prejudice", according to both reports.

For computing teachers, Powell set no date and described no new rule. The National Centre for Computing Education said in a September blog post that the department is expected to publish a draft refreshed computing curriculum this autumn term. It said the new curriculum is expected to be taught from September 2028. It also told teachers there is no need to make speculative changes now.

Powell did not say what more practical assessment would look like in a computing GCSE. She did not say whether it would mean on-screen programming tasks, more coursework or something else. It is not yet known when the draft curriculum will be published.

The parliamentary record, which is in Dutch, shows 117 of the 150 seats in favour on a show of hands. The VVD, BBB, FVD and one independent MP voted against. The motion asks the cabinet to include the ban in a digitalisation plan for education that it is preparing.

Judith Tielen, the state secretary for education, advised against the motion in the debate on 29 October. She said she understood the concern but needed more time. She asked Volt leader Laurens Dassen to hold the motion back until her plan is ready in December. He refused.

Volt and the CDA said children should learn to read, write and do arithmetic before they use AI, the Dutch-language Hart van Nederland reported. The parties cited PISA research and said pupils who use generative AI score 20 points lower in science on average. They still encourage teachers to use AI to improve lessons.

The PO-Raad, which represents primary schools, said early research shows pupils learn less when they use generative AI heavily, DutchNews reported. It also said a total ban came out of the blue and that it would prefer an agreement with the ministry.

For Dutch parents and teachers, nothing changes yet. A motion is a request to the cabinet and does not change the law.

DutchNews said new learning goals due in 2027 require primary pupils to explore AI and learn to use it responsibly. It said the motion does not rule out teaching children about AI. Dassen told MPs the motion treats generative AI and educational AI as different things.

It is not yet known whether the cabinet will carry out the ban. It is also not known how a ban would define generative AI or how schools would apply it.

The report says 37 institutions joined in September 2025 and 28 stayed active to the main evaluation. They used TeacherMatic and Inspera Graide. A third product, KEATH.ai, withdrew. The report follows interim posts in May.

Jisc says AI was most useful for repeatable tasks where staff kept responsibility for academic judgement, such as drafting and structuring feedback. Workload savings were possible but conditional. They were clearest with digital submissions, explicit criteria and large batches of similar work. In long or interpretive tasks, staff could lose the saving to checking and rewriting.

Participants reported that AI could be too generous to weak work and did not always recognise exceptional work. Some tools struggled to read handwritten, visual or very long work in full.

For students, Jisc calls its evidence tentative because it gathered less direct feedback from them. Some accepted AI feedback if it was faster or clearer. Others raised concerns about fairness and trust. Views were more positive where lecturers explained why AI was being tested and how students could challenge the process.

In one effective approach for lecturers, the marker reads the work and makes notes before seeing any AI output, so the tool does not steer the judgement. Most activity in the pilot was formative, meaning work used for development and not for high-stakes results. Jisc says use in final marks needs a higher level of evidence and oversight.

Jisc calls the findings indicative, because this was a practice-based pilot and not a controlled comparison. The report gives no overall figure for hours saved. Case studies on custom AI assistants are due later in 2026.

Chris Phillips, who leads education at Google, set out the change in a blog post. It covers Gemini, Gemini Notebook, Gemini in Classroom and Google Search. Each will stay off for K-12 student accounts until a school administrator turns it on. K-12 is the American term for the school years from kindergarten to the end of high school.

The news agency AFP added detail in a report published by Phys.org. Google told AFP the controls will be rolled out worldwide, beginning in the United States, but that no start date has been set. AFP also reported that schools will not be able to give pupils Google Search without its AI features. It said two US senators wrote to Google last month to ask why AI features were the default on school devices.

For teachers, this means pupils will lose Gemini in lessons unless the school's administrator turns it back on. A school that wants pupils to use Search must accept its AI features too, according to AFP. The same Google post said Guided Learning is now available worldwide in Google Classroom. It is a Gemini mode that asks questions and breaks a problem into steps instead of giving the answer.

Google said it will give schools more details in the coming months, before the change starts. It has not said when the controls will reach schools in the UK. It has also not said what happens to accounts where administrators have already turned Gemini on.

The guidance, from the Department for Digital, Culture, Media and Sport, says the scholarship pays £4,000 for each year of full-time study at a UK university, for up to four years. Scholars also get summer placements of about eight weeks in a technology field, plus mentoring and careers support. The scheme focuses on six areas, including AI, cyber security and semiconductors.

Applicants can study any subject. School and college students must expect at least 120 UCAS tariff points (BBB at A level) and plan to start a full-time degree in September 2027. Current undergraduates must be on track for at least a 2:1 and have two years of full-time study left from September 2027. Applicants must qualify as UK home students.

Selection has three online stages: a form, an assessment and a video interview. The department says the process tests how candidates learn and solve problems, not what they already know about technology. It says it particularly welcomes women and students from lower-income backgrounds.

For a sixth former or student interested in tech, the deadline is 15 January 2027. Coding knowledge, tech qualifications and industry experience are not required. Most other UK government scholarships cannot be held at the same time.

The scholarship is part of TechFirst, the government's tech skills programme for young people. The guidance does not say how many scholarships will be awarded.

The course is called Learn AI Coding. Boot.dev co-founder Lane Wagner said in the company's November newsletter that it was made with Dax Raad, a co-founder of OpenCode. OpenCode is a coding agent, a tool that uses an AI model to write and change code when asked. Wagner wrote that learners use the agent to build an interpreter for a small language called SlopLang.

The course page says the course has 13 chapters and 95 lessons and takes about 20 hours. The chapters cover models, prompting, agents, context, rules, skills and plugins. The page describes the project as a Python interpreter. It lists the Boot.dev Team as author and Raad as a contributor.

Boot.dev says the course is for people who can already write and debug simple code. Wagner wrote that it is not a "vibe coding" course, in which people let AI write code they do not check. Learners do not need a paid AI subscription, because free models are available through OpenCode Console.

The course page says anyone can make a free account and try a few chapters. After that, learners without a paid membership can only read the content. The course sits in the Backend, DevOps and Data Analyst paths, just before the final project.

Neither page gives the price of a membership or the exact day the course went live. Wagner said courses on Terraform, Bash scripting and Lua are coming next, without dates.

The announcement covers courses starting between 1 January and 31 July 2027. The LLE replaces the old undergraduate loan system for students funded by Student Finance England. Someone who has not had student finance before can borrow up to £39,160 for tuition. That is about 480 credits, where a full-time degree year is usually 120.

The main change is that loans can also pay for parts of a course. The Department for Education's overview lists 10 subject groups where single degree modules can be funded. Computing is one of them, along with engineering and mathematical sciences. A module must be worth at least 30 credits and come from an existing full course at a provider registered with the Office for Students.

The opening date moved once. The overview said in September that applications would open in late October rather than in September. Providers that want to offer funded modules from September 2027 can apply to the department until 30 November 2026.

For an adult who wants to retrain in computing, this means a loan for a single module of a degree from January 2027, if a local provider offers one. Applicants need a GOV.UK One Login account. Tuition loans are open to people under 60 when the course starts. The University of Greenwich notes that student funding in Wales, Scotland and Northern Ireland has not changed.

The government has not said how many computing modules will be available in January. Applications for courses starting in the 2027 to 2028 academic year open in spring 2027.

The paper, by Grace Kerr, Aarushi Parasrampuria, Faith Culas and Kelly Blincoe, is on arXiv, a public archive of research papers. Twenty students who knew Java used Copilot to write, debug and refactor code while saying their thoughts aloud. Refactoring means rewriting code to make it clearer without changing what it does. The sessions took place in July and August 2025.

The team used GenderMag, a method for finding software features that work less well for some users. It compares people on five traits, including motivation, attitude to risk and learning style. Some students who wanted to know why Copilot made a suggestion found the explanations long and in a separate panel. Some students blamed themselves when they could not find features such as Chat and Agent mode.

The authors found that Copilot helped students finish tasks but gave little support for learning. Some students accepted code while unsure and then removed it with Undo. Others felt pushed towards one complete solution. Verified students and teachers can get a plan called GitHub Copilot Student free, according to GitHub Education.

The authors suggest short explanations next to each suggestion and optional learning modes with links to read more. They say these ideas have not been tested and that the sample was small and from one university. Their findings reflect the version of Copilot used at the time. It is not known whether later versions behave the same way.

The State Educational Technology Directors Association (SETDA), which represents these officials, said in its press release that 79% named AI a top priority. That is up from 68% in 2025. It said 23% of states now have a law that requires AI guidance for schools, up from 4% in 2024.

The report draws on 73 survey responses from officials in 45 states and one territory. Most figures are shares of those respondents, not counts of states.

Rules for buying AI are getting stricter. SETDA said 53% of states have published guidance on buying AI tools. Of that guidance, 79% now sets evidence standards, up from 45% last year. But 68% of state leaders said it is hard to check vendors' claims about AI products for themselves.

EdWeek Market Brief reported that the report describes a shift from giving pupils devices to proving that the technology works. SETDA said 52% of respondents reported that their state collects no data on how its edtech is used or whether it works. It said 57% reported a strict statewide ban on pupils' phones, up from 22% in 2025.

For teachers, the fastest-growing need in the survey is training. SETDA said 62% of respondents named professional learning as a top unmet need, up from 47% in 2025. These answers come from state officials, not from teachers.

The survey does not measure what happens in classrooms or whether AI tools improve learning. The press release does not name the states that have laws requiring AI guidance.

The regulator's rapid assessment names photo ID matching, banking checks, digital ID services and facial age estimation. The government asked for it before setting the rules for its ban on social media for under-16s.

Ofcom said a face scan can be wrong by a few years at 16. People who look close to the limit should get a second check. Its example is a second check for anyone who looks under 19.

Ofcom said 16 and 17 year olds are more likely than adults to need photo ID, because every method at 16 uses it or falls back on it. It cited an estimate, from a survey in schools that was not representative, that about 150,000 of them have neither a passport nor a driving licence. The law firm Osborne Clarke noted that Ofcom found credit card, mobile phone and email checks not good enough at 16.

Online safety minister Ian Murray told MPs on 2 November that the first regulations will be laid before Parliament's Christmas recess. The ban would start in spring 2027. He said the regulations will not name platforms.

The government's fact sheet says the UK will follow Australia's model, which covered YouTube among other platforms. It also says dedicated educational platforms that support schoolwork and learning will not be covered.

Pupils and parents need do nothing yet, the government says. A 16 or 17 year old without a passport or driving licence can apply for a PASS proof of age card. Ofcom said some are free for this age group.

Which platforms the ban covers is not yet settled. The government has not said how pupils under 16 would reach learning videos on platforms such as YouTube. Ofcom will consult on its guidance once the regulations are laid.

The School and College Voice survey ran from 30 June to 13 July. In it, 21% of secondary leaders said pupils hand in phones or leave them somewhere they cannot reach. That compares with 9% in May 2025. Most secondary leaders (70%) still let pupils bring phones if they stay out of sight and unused, down from 75%.

Statutory guidance says schools should be phone-free by default. Schools should begin following it from 1 September 2026.

Among secondary leaders with a phone policy, 38% said they had updated it and 46% said they already met the rules. Another 11% planned changes and 4% did not. Among primary leaders with a policy, 16% did not plan to change.

The survey also shows a gap between leaders and classroom teachers. Rules were followed most or all of the time according to 92% of secondary leaders and 78% of secondary teachers. Most secondary teachers (65%) said they spend under 15 minutes a day enforcing the rules. Another 20% said they spend 15 minutes or more.

For parents, the guidance expects that pupils will not have access to phones throughout the school day. That includes lessons, breaks and lunchtime. Headteachers decide how to achieve this, so the method will differ between schools.

The survey counts what leaders and teachers said, from 512 secondary leaders and 508 secondary teachers with a policy. It is not a check of schools. It was taken before the guidance applied, so it does not show what schools have done since September.

The preliminary report from its research centre groups computer and information science together. It found falls of 12.3% at two-year colleges and 10.4% among graduate students. Across all subjects, total enrolment rose 0.2% and undergraduate enrolment rose 0.6%.

The figures cover 7.92 million enrolments reported by 24 September 2026, the Clearinghouse said in its statement. They come from 46.0% of the US colleges and universities that send it data, so they may change. Bloomberg, in a report published by The Star, said undergraduate computer science courses have lost 18.4% of their students over two years. Graduate courses lost 27.1% over the same period.

For students thinking about a computer science degree, Higher Ed Dive reported a related figure from the Federal Reserve Bank of New York. It put unemployment among recent computer science graduates at 7% in 2024. Bloomberg said the preliminary data do not include information on international students. The final report, due in January 2027, will include it.

Neither the report nor the Clearinghouse's statement gives a reason for the fall. Matthew Holsapple, the centre's senior director of research, told Bloomberg: "Those are just remarkable declines for a field that until recently was one of the fastest growing areas that we’d seen in higher education." It is not yet known whether the final figures will show the same size of fall.

The governor's office said companies may not offer these features to anyone under 18. They also may not offer them to any user whose age they have not confirmed.

Banned features include pretending to be a real person or to have human emotions. Chatbots may not claim a personal, romantic or authority relationship with a child. They may not promote self-harm, tell children to keep secrets or produce sexually explicit content.

Companies must offer more than one way to check a user's age. At least one must not depend only on a government ID. Age data must be deleted after use. The state attorney general can seek penalties of up to $25,000 for each violation, the office said.

AP reported that the law lists 12 definitions of unsafe companion features. Justine Gluck of the Future of Privacy Forum, a nonprofit, told AP that some categories are broadly worded or undefined. Law Commentary reported that a state business group had warned an earlier version could affect educational tools.

For parents and teachers, the law does not ban children in New York from using chatbots. It bans specific features. Melinda Person, president of the New York State United Teachers union, welcomed it in the governor's announcement.

The law applies only in New York State and changes no rule in the UK.

The start date is not settled. AP said the law takes effect on 1 January. Law Commentary reported that the office of Senator Kristen Gonzalez, who sponsored the bill, said an amendment moves the date to 1 June 2027. It is not known whether AI companies will challenge the law in court.

In a blog post looking back on this year, the foundation said there will be two events in the UK. Coolest Projects UK Bradford will take place on 15 May 2027 at the Life Centre in Bradford. Coolest Projects UK Cardiff will follow on 22 May 2027 at the Wales Millennium Centre.

The foundation also said online entries will open in January. Its Take part page gives the date as 13 January 2027 and lists both UK events as free. The online celebration will be streamed on 23 June 2027.

Pupils up to the age of 18 anywhere in the world can enter online, according to the foundation's help page. Its website says learners aged 13 or over can register a project themselves, but an adult must register for anyone younger. It lists seven categories, including Scratch, games, web, hardware and artificial intelligence. Every entry is checked before it appears in the online gallery or gets a place at an event.

Teachers and coding clubs can sign up to the foundation's newsletter to hear when entries open. The foundation said more than 15,000 young people from 40 countries shared over 4,500 projects online in 2026. It has not said when entries for Bradford and Cardiff will close and said the full calendar of events will follow soon.

In a post on its Skills Hub blog, Microsoft said the badge is based on how a person uses GitHub Copilot, its AI coding assistant. It measures how broadly and deeply someone uses Copilot on GitHub.com, in code editors, at the command line and on mobile. The framework has five levels, from Beginner to Expert. A person qualifies for the badge on reaching the third level, Proficient.

The Register reported on 5 November that Microsoft had published criteria only for the first three levels. It said the badge does not replace Microsoft Certifications or Microsoft Applied Skills. Microsoft described it as an extra signal of how skills show up in daily use of a product.

For new programmers, Microsoft said the badge costs nothing and needs no exam or study guide. A learner signs in with a personal GitHub account and opts in. Microsoft said raw code and prompts are never reviewed. A person who qualifies can claim the badge on Credly, a site that issues digital certificates, then share it on LinkedIn.

Accounts that an employer manages are not supported at launch, Microsoft said. It has not said when the Advanced and Expert levels will open. Neither source says how employers will treat the badge when they hire.

Code Classroom is the new name for the Code Editor for Education, the foundation said in a blog post. In the new Blocks editor, pupils drag and drop blocks to make games, stories and animations without typing code. The tool already had a Python editor and a web editor for HTML and CSS. All three now sit in one place.

The foundation said the Blocks editor is based on the Scratch editor, which is open source. Open source means anyone can read and reuse the program's code. The Scratch Foundation publishes the code for the Scratch 3.0 editor's interface on GitHub under the AGPL-3.0 licence.

Teachers can set up a free school account on the foundation's website. The tool runs in a web browser, so schools do not need to install anything. The foundation said there are no subscriptions or hidden costs and that teachers can follow each learner's progress.

Teachers can also import classes from Google Classroom. Pupils can then sign in with their Google accounts.

The foundation said more than 4,500 schools already use the tool to teach Python and web development. Bartholomew Owuor, a computer science teacher at Cathsam School in Kenya, said the blocks let him include learners of different ages. The post does not say which age groups the Blocks editor is meant for. It also does not say how the tool fits the computing curriculum in England.

Informatics is the word many European countries use for computer science. The announcement said the festival brings schools, colleges, universities and employers together to improve gender equality in computing. The festival receives a prize of €5,000.

The festival began in 2020 and is now a registered charity, the announcement said. In 2025 it ran 24 events across Scotland, from the Borders to the Highlands. They reached more than 2,000 people. Nearly half came from Scotland's most deprived communities.

Informatics Europe also noted that the first Ada London Festival launched in 2026. Festival director Matthew Barr said: "The prize money will help us grow still further".

The festival's website says the 2026 festival runs from 18 September to 5 November. It lists online and in-person events for teachers, learners of all ages and parents, including a teachers' evening forum and a hackathon in Dundee. Teachers can be nominated for its Inspiring Educator Award 2026. The site says girls made up 19% of entrants for Higher Computing Science and 22% for National 5 in Scotland in 2023.

The award committee, chaired by Professor Ana Belén Gil González of the University of Salamanca, judged nominations on impact and reach, sustainability, how easily others could copy them and originality. This is the 11th year of the award. SAP is its new sponsor.

The announcement does not say when the prize will be presented or how many nominations were received.

EU Code Week invites people of all ages to explore coding. Its about page says it began in 2013, is run by volunteers and is supported by the European Commission. The average participant is 11 years old. The main event lasts two weeks each October and ran from 10 to 25 October in 2026.

In a post on the EU Code Week blog, the hub said over 28,000 activities were registered in Italy over the same two years. That is up from more than 24,600 when the figures were first uploaded. The Italian version of the post says the count of students rose from nearly one million to more than 1,150,000. The hub is run by Fondazione LINKS in Turin and has been the national hub since 2024.

The post was written after the Italian newspaper La Stampa reported on Code Week on 9 October 2026. That article gave a lower figure of 950,000 participants, because it was published before the count was updated. The hub said the number of its Leading Teachers doubled from 35 to 70.

The about page says Code Week is now a year-round movement, so schools are not limited to the October fortnight. The post does not say how students were counted or whether a student who joined several activities was counted more than once. It also gives no figures for each region or year.

The Digital and Technologies Jobs Plan says around 800,000 digital and technology roles will need filling over the next decade. The government and industry aim for around 200,000 entry-level jobs over that time, including apprenticeships. Government and employers commit to about 20,000 a year through a single "Skills Deal".

The plan says many young people who took a computing course or degree find it hard to get in. In cyber security, the share of job adverts open to people with less than a year of experience fell from 25% in 2022 to 16% in 2025. In 2025, 64% of those adverts asked for two to six years.

For learners, the plan lists one change with a date. From April 2027, anyone who completes a digital T Level, V Level or apprenticeship can be recognised by BCS against a professional standard at no cost. BCS confirmed this in an article on its site. The plan notes that T Levels, V Levels and several other named routes are for England only.

A related TechFirst AI Apprenticeship Pathway is being piloted in North West England. It trains young people in AI and digital skills, then helps them apply for Level 3 digital apprenticeships. The plan calls the 200,000 figure an ambition shared by government and industry, not a forecast. It says progress will be reported each year.

The foundation built Ada with the University of Cambridge for students aged 14 to 19. It said in a blog post that the new lessons are the first of several releases paid for by the Department for Digital, Culture, Media and Sport. Later releases will cover topics such as AI and cybersecurity, plus self-paced projects and careers case studies. The foundation said everything stays free.

The work is part of TechFirst, a government tech skills programme. On the same day, the government said in a press release that every secondary school can now join it. The government aims to reach 1 million students by 2028, a year earlier than planned. Employers including Microsoft, IBM and Arm will offer 10,000 work experience places over three years.

Teachers and pupils in the UK can try the first new material on the Ada website now. A teacher who wants the wider programme signs the school up by emailing a regional delivery partner. The release lists an address for each English region and for Wales, Scotland and Northern Ireland. The foundation said these partners run practical technology activities in schools.

The release also said Ada will replace "the Issac platform currently funded by the DfE" from the end of the 2026 to 2027 school year. It did not say what happens to pupils' work stored there. Neither source said when the later Ada releases will arrive.

A boot camp is a short, paid course that trains people for a job. Sarah Butrymowicz, the outlet's investigations editor, wrote in a follow-up article that IronCircle promised students a voucher worth about $400. The voucher paid for CompTIA Security+, an industry exam that tests basic cybersecurity skills. Two former employees told her the company stopped handing them out in summer 2025.

In January 2026 one student asked about his voucher. He was told the programme now recommended a new credential that IronCircle had created. Steven Ostrowski of CompTIA said by email that he did not know that credential. "Some of what students learn in the Iron Circle course may help with our exam, but probably not enough to pass our test," he said.

For learners, the article separates a known exam such as Security+ from a certificate a course provider makes itself. It said some credentials are required by law but many more are set by industries themselves. Credential Engine counted 1,850,034 unique credentials in the United States in a report of 9 December 2025. Former IronCircle staff said vouchers went only to students who kept asking.

Few of the former students Butrymowicz interviewed took and passed the CompTIA exam. IronCircle closed in June and did not answer a dozen requests for comment before then. The article did not say how many students received the new credential.

QAA said the principles give colleges and universities a shared reference point for one question: whether an assessment shows what a student has learned. It wrote them with an advisory group of researchers, university leaders, practitioners and student representatives. They are meant to apply from further education to doctoral level.

The draft text says summative assessment, the work that counts towards a grade, should include some points where a student's own attainment is checked under conditions in which AI use can be controlled or verified. Not every assessment needs this, it says. A course can plan these points across a whole programme. Examples include questioning, practical work, observed work, placements, simulations and presentations.

The draft says total bans on AI are hard to enforce. It says current detection tools are widely reported to have significant rates of false positives and false negatives, so providers should not rely on detection software alone as a main safeguard. It also calls it "insufficient, and unfair" to set AI rules that a provider has no practical means of checking.

For students and lecturers, nothing changes yet. QAA says the text has not been agreed or endorsed by its advisory group, by QAA or by any other body. Students and staff can send comments through a form on the draft page. The draft says limits on AI should not extend to assistive technology or reasonable adjustments.

QAA says a final version will follow later this year with shared definitions of assessment terms and recommendations. Its pages give no closing date for feedback. They also do not say whether the Office for Students, the regulator in England, will refer to the principles.

TechFirst is the government's tech skills programme. The guidance says it is "giving 1 million young people the chance to build their skills, get confident with AI and explore future careers in tech." Schools and colleges take part by contacting a regional delivery partner and can apply to become a Bronze, Silver or Gold TechFirst School. The main TechFirst page already had a schools section, last changed on 30 September.

The main online resource is Ada Computer Science, a free platform with coding projects and revision content for students aged 11 to 19. TechFirst supports it through a partnership with the Raspberry Pi Foundation. The guidance says new resources for pupils aged 11 to 14 are being written.

Teachers must register pupils for the TechFirst Girls Competition, run by IBM for girls aged 12 to 13, by 10 November. The competition runs from 11 to 20 November. It replaces the CyberFirst Girls Competition, which ran from 2016 to 2025. New challenges cover AI and quantum computing.

The UK Bebras Challenge runs from 9 to 20 November for pupils aged 6 to 19. It tests computational thinking, which means breaking a problem into steps a computer could follow. Both challenges are free and need no coding experience. The guidance does not say how many schools have joined TechFirst so far.

Three of the authors work at Aalto University in Finland and three at universities in the United States. The paper covers an online introductory course from September 2025 to June 2026. After a wrong submission, a large language model, the kind of AI behind chatbots, wrote a hint for the student. The course used o3-mini, then GPT-5-mini, then GPT-5.5.

The authors said feedback went with better next attempts in several comparisons, but the pattern varied. Longer messages went with fewer correct next submissions, more attempts and longer gaps before students tried again. Students who had feedback earlier in a course part were less likely to solve the related final-project step. From January to May 2026 the course had 525 students who made 83,374 submissions.

For teachers who add AI hints to programming exercises, the authors suggest testing two changes. One is a limit on the length of each message. The other is letting students choose between a short hint and a fuller explanation. They said both ideas need proper trials.

The paper is listed for 6 November 2026 at Koli Calling, a computing education conference in Finland, the conference schedule shows. The study measured later performance, not learning itself. It is not known whether shorter hints would help students learn more.

The working paper has not been peer reviewed. It covered 2,379 undergraduates and 30 instructors in autumn 2025. Instructors were randomly assigned to add the tutor to their courses or not. The tutor was built on GPT-4o and drew on each course's materials.

The clearest result came from 1,353 students in sections of the same course. Their final grades were 4.12 points lower out of 100, which is 0.37 standard deviations. Across all 2,379 students, the authors found no significant evidence of lower grades.

Recorded participation in course activities fell sharply in both samples. Estimated grade losses for first-generation students were more than double those of other students.

Only about 15% of students offered the tutor ever used it, so the authors say the effect may reflect wider AI use prompted by the tutor. Students in both groups could already use ChatGPT and Gemini. About 74% of requests sought information, explanations or solutions. Fewer than 1% asked for feedback on a student's own work.

The Hechinger Report said lead author Jing Liu does not conclude that colleges should avoid AI. A separate peer-reviewed trial in Nature Human Behaviour on 26 October tested 487 management consultants. ChatGPT helped them score higher on data science tasks, but the authors found no evidence they did better on technical questions without it afterwards.

For students and lecturers, the authors say the result applies to this tool as set up in this trial, not to all AI tutors. Only two instructors ended the term with the mode that guides students rather than giving answers. Grades were not an independent test of learning. Maryland Today said 40 courses have the tools and their instructors have received guidance. It does not say whether the university will offer the tool to every course.

Paul Waugh gave the update for the department in a written answer to the MP Dr Ben Spencer. He said the tools are being designed for year 9 and 10 pupils. This involves closely supervised research with around 30 partner schools. Tools that pass safety and teaching tests will move to trials through the EdTech Testbed Programme, run with the Education Endowment Foundation.

Spencer asked when trials would begin, how many pupils would take part and whether the design would be randomised. The answer did not give a date or a number. It said study plans will be published and will set out the method.

The department said in a press release in January that trials would begin later this year and tools would reach schools by the end of 2027. Schools Week reported in July that eight firms are working on the tools, after Google left the supplier list.

Concerns were raised in Parliament days later. Meg Moss of the National Counselling and Psychotherapy Society told the education committee on 20 October that she worries about the safeguarding risks of AI tutors.

For parents and teachers, the answer means the tools are still in supervised research in a small group of schools. It describes no way for other schools to use them yet. It does not say when trials will start or whether the end of 2027 target still stands.

Both sources for this story are in German. The translations here are our own. The ministers meet as the Bildungsministerkonferenz, part of the Kultusministerkonferenz (KMK). Its press release says pupils should understand AI and use it creatively and for a purpose. They should also judge its results critically and weigh its chances and risks.

The framework keeps its six existing areas and adds AI to them. Examples include spotting bias in AI output and labelling work made or edited with AI. Pupils should also learn how to deal with deepfakes. The framework itself says it follows the AI literacy framework of the OECD and the European Commission closely. Its area on algorithms covers the central ideas of machine learning and their limits.

For teachers in Germany, the press release lists support that is being prepared. It includes an interactive overview of the six areas and 22 reflection cards for schools. There will be a digital version of the cards and an AI skills guide in AIS.chat, a chatbot shared by the states. A KMK working group led by North Rhine-Westphalia is writing the material.

The framework is already on the KMK website. The support material will be published there after it has been presented at events. The press release gives no date for those events.

Ofqual regulates exams and qualifications in England. Bauckham spoke at the Confederation of School Trusts annual conference in Birmingham. In the speech Ofqual published, he said extended writing coursework is open to cheating. He also said research shows that students who use AI as a shortcut achieve less than those who think for themselves.

Bauckham separated machine marking of questions with clear right answers from AI marking of essays and other long answers. He said AI marking "opens up real ‘routes to harm’" because AI makes things up and reflects biases. Tes reported that he warned against "undiscriminating use of AI" in schools.

On-screen exams already reach some computing pupils. Schools Week reported in May 2026 that pupils sat the second paper of Pearson Edexcel's computer science GCSE on screen. Bauckham said some exam boards want Ofqual to move faster. He added that Norway, Sweden and Denmark had paused or reversed moves to on-screen exams.

The speech set no new rule or deadline for pupils or teachers. Bauckham said Ofqual is looking at the digital divide and school readiness as it decides how to regulate on-screen assessment. He told the audience to "Expect to hear more in due course." Ofqual has not said when it will decide on on-screen exams or what limits it will set on AI marking.

The post by Michael Webb updates his 2025 advice. He says commercial detectors wrongly flagged 0% to 4% of human writing across four studies. He warns that the human texts tested were arguably good writing. They did not include weak first-year essays or heavily proofread work by second-language writers.

Webb cites a preprint, which is not peer reviewed. He says it found that Pangram, the tool he calls the best performing at present, caught 34% of text from GPT 5.4 soon after that model's release. A May update fixed this. Adversarial rewriting in the same paper cut its recall from 94.7% to 24.1%.

Webb cites a study with 30 texts in each of five conditions. Human reviewers scored 19% overall, against 20% for guessing. His advice is not to ask staff to spot AI use as they mark.

Some awarding bodies still require a detector. Webb says there are none Jisc can honestly recommend that suit small cohorts, are reliable enough and have passed data protection checks. He asks colleges in that position to contact Jisc.

For students, the post says a flag is the tool's judgement and not an established fact. It says a flag does not show whether the use was permitted. For lecturers, it calls assessment redesign the only viable way forward and asks that any new measure be reviewed for equality effects.

Webb calls wearables the next challenge for in-person exams. Ofqual said phones and smart devices made up 2,225 malpractice cases in summer 2025. That was 44.3% of student cases in school exams, not university assessment. The post is advice, not a rule. Webb says he cannot find whether UK institutions qualify for access to one AI text watermark checker.

ISE's summary of its Student Recruitment Report 2026 compares that with 85% in 2025 and 78% in 2024. The figure counts employers, not candidates, so it does not show how many applicants broke the rules. The most common forms were unauthorised AI use in interviews and online assessments. Wonkhe reported that 66% of employers identified unauthorised AI use in interviews.

ISE says employers received 141 applications per graduate role and 96 per apprenticeship. It says 77% are reviewing or redesigning their recruitment. Measures include more personal questions, time limits and randomised questions. Some employers limit how many roles a person can apply for or close campaigns early.

Graduate hiring rose 4% after an 8% fall the year before. A later ISE article says growth falls to 0.8% without a few large employers and that the typical employer's growth was zero. It says Digital and IT grew fastest in proportion, together with the built environment, but gives no figure for either. Employers forecast 7% growth in graduate hiring for 2026/27.

For students applying now, ISE says 82% of employers give guidance on acceptable AI use, up from 55% in 2025. Applicants can look for that guidance before using AI on a test or in an interview. ISE says 16% of employers use AI in their own recruitment and 29% plan to start within a year.

The survey covers 155 employers and more than 1.7 million applications. ISE says it shows larger employers best and says less about small firms. The full report is for members and paying customers. The public summaries give no figures for software or computing roles.

The checklist reminds schools that the requirements of the Data (Use and Access) Act 2025 are in force. It says the Act allows wider use of automated decision-making if safeguards are in place. An automated decision is one made by a computer system without a person. Schools are told to find any systems that do this and make sure a person can review and change the outcome.

The checklist also tells schools to carry out data protection impact assessments where required, particularly for higher-risk use of AI tools, EdTech and monitoring technology. It says privacy information for children should use clear, age-appropriate language.

The Information Commissioner's Office, the UK's data regulator, says the Act opens up more lawful reasons for making significant automated decisions, as long as safeguards apply. It also says organisations must acknowledge complaints within 30 days. The department's checklist repeats this for schools.

For parents and pupils, the checklist says people must be told when an automated decision is made. They must be able to challenge it and ask for a human review. People should normally complain to the school before going to the regulator.

Schools can now pause the one-month limit on a request for personal data if they need clarification. They need only make reasonable and proportionate searches.

The checklist gives no examples of which school systems count as automated decision-making. It sets no date for finishing the review. It does not say how many schools use such systems.

A boot camp is a short, paid course that promises to prepare people for a job. The Hechinger Report said that more than 20 universities offered IronCircle courses under their own names. The company charged students as much as $18,900. Reporters interviewed more than 35 former students and staff and reviewed university records.

Three universities had complete job records. At those, only 35 percent of graduates were recorded as finding work within six months. At the University of Central Florida, 26 of 325 students were recorded as finding a job in that time. IronCircle did not reply to requests for comment.

Before paying for a boot camp, learners can check the things the report found missing at IronCircle. The company was not accredited itself and students who needed help paying had to take out private loans. The enrolment agreements Hechinger reviewed were between the student and the company, not the university. One student who asked for job placement figures was told the company did not track them.

The US Department of Education said in January 2025 that such partners must not give misleading information about graduates' employability. QuickStart, which bought some of IronCircle's assets, said it would let some students transfer. Experts told the Hechinger Report that students will be at the bottom of the list in the liquidation. It is not known how much money, if any, they will get back.

The signatories include GPs, paediatricians and speech and language therapists, Tes reported. The letter to Lucy Powell and Yvette Cooper calls the lack of child-focused health scrutiny "a serious safeguarding concern". It says about £900m a year is spent on edtech in English schools.

EdTech Innovation Hub reported that the letter covers school tablets, laptops, interactive whiteboards and headphones. It also covers the learning platforms used on them. The clinicians want health considered when schools evaluate and buy these products. They also want evidence-based guidance on daily screen exposure.

Health Professionals for Safer Screens announced the letter. The campaign group SafeScreens supports it.

The letter cites analysis that almost one million children spend nearly every lesson on screens, LBC reported. These are the campaigners' figures and the reports do not show that they have been independently checked. The reports also do not show that school screens cause the health problems the letter lists.

For schools and parents, no rule has changed. A government spokesperson told Tes that technology used well can be a "powerful tool" and that a consultation on AI and edtech safety would take place.

There has been one development since. Tes reported on 3 November that early education minister Paul Waugh told a Commons committee the government is testing edtech products with teachers. He said it is also looking at secondary legislation to inform a statutory code of practice from the Information Commissioner's Office.

The reports give no date for the consultation. It is not known whether physical health will be part of any code.

The roundtable met on 7 September 2026 and was chaired by Professor Dame Angela McLean, the Government Chief Scientific Adviser. The notes give the combined views of the experts present. They are not a statement of government policy. Those present included Peter Kemp of King's College London and Sue Black of Durham University.

The notes cite research that science ambitions differ by gender from the age of 10. They say girls are particularly underrepresented in A level computer science, physics and mathematics. They also say a relative lack of Ofsted reporting on computing could indicate that schools give the subject less priority. Durham University's TechUP programme has retrained about 800 women since 2018, the notes add.

For teachers, the notes list approaches with some evidence behind them. Girls respond well to group work where they can discuss and compare ideas. The notes say gender equity work across a whole school is effective, rather than work aimed at one subject. A randomised trial in more than 130 schools raised the share of girls choosing physics.

For adults, TechWorks reported that the government's TechFirst Women's Programme launched on the same day. It aims to place 300 women in paid tech jobs by March 2027. Women can register interest at techfirstwomen.com. The government has not said whether it will act on the advice in the notes.

The European Commission said on its digital strategy site that the event was streamed from Brussels from 14:00 to 16:00. It said Code Week 2026 runs from 10 to 25 October with activities across Europe. The theme is "Building Resilience, Connecting & Empowering Communities in the Age of AI".

A summary on the Code Week blog, published on 14 October, says one panel asked why coding still matters as AI tools spread. Speakers from Bebras, SAP and Cisco Networking Academy took part. They said computational thinking and the habit of checking results that AI produces are becoming more valuable. Computational thinking means breaking a problem into small steps that a computer could follow.

For teachers, the blog lists several ways to join before 25 October. They can run a coding workshop, introduce pupils to AI or join a local activity. The launch also presented four free Minecraft Education lessons on AI literacy that come with training resources for teachers. A new Velocity Prize is described as a global Minecraft Education challenge for students aged 13 to 18.

Henna Virkkunen, an Executive Vice-President of the European Commission, sent a video message. So did Roberta Metsola, President of the European Parliament. The blog post does not give a closing date for the Velocity Prize. Neither page says how many schools or learners are expected to take part this year.

The paper, by Mykhailo Kopytskyi, Jianhua Yang and Mir Seyedebrahimi, describes a tutor called SocraticRepair. A Socratic tutor asks questions that lead the learner to find the problem. The team further trained a small open model, Qwen2.5-Coder-7B-Instruct, on 1213 tutoring dialogues about Python. Other AI models wrote these dialogues from buggy student programs in four programming datasets.

The team compared it with GPT-4o, a general chatbot model, on 53 programming problems it had not seen. According to the full text, SocraticRepair gave away the answer too early in 24 of 1,276 turns (1.88%). GPT-4o did so in 59 of 794 turns (7.43%). Across whole conversations, the figures were 12.6% and 33.3%.

For learners who ask a chatbot for help with code, the paper says general models tend to reveal the solution too early. The Raspberry Pi Foundation makes a similar point in a lesson unit for ages 13 to 16, made with Google DeepMind. The foundation says AI models usually give the answer by default. Its five lessons teach pupils to write prompts that get the tool to guide or challenge them instead.

GPT-4o scored higher on technical correctness. The lesson plans SocraticRepair wrote for itself often had technical errors. The students in the tests were simulated by another AI model, which the authors say may not reflect how real learners would use the tutor. The paper does not say whether the tutor will be released.

The magazine said they follow the University of Southampton, which announced the same decision earlier this year. The concern began with proposed new Turnitin licence terms on processing student submissions with AI tools. The University of Aberdeen said Turnitin has postponed those terms from 7 July 2026 to September 2027. Aberdeen said it is reviewing its own use of Turnitin this year.

Newcastle said its education committee decided in September to begin a phased withdrawal. It cited other plagiarism tools it already has, the student experience and the potential use of staff and student data for AI from next year.

Westminster gave different reasons. It said running two submission systems confuses students. It also said similarity checking is "largely ineffective against the most significant current integrity risk", because generative AI writes original text. It said AI detection carries a risk of false positives, particularly for non-native English speakers.

Turnitin told Times Higher Education that it does not train a generative AI model. It said it trains a classifier that predicts whether text was written by AI. Thomas Lancaster of Imperial College London told the magazine that money is the deeper reason for universities leaving.

Westminster students will submit all coursework through Blackboard Assignment from February 2027. They can see past Turnitin work only until 31 July 2027, so the university tells them to download anything they want to keep. Newcastle staff are moving to Canvas assignments.

Newcastle has not yet chosen a replacement similarity checker. The final wording of Turnitin's licence is not known. Times Higher Education said several more universities are considering their contracts but gave no total.

In its announcement, the department said the options would mix practical technical learning with core subjects such as English and maths. It gave coding, construction and manufacturing as examples of technical learning. The poll covered 500 young people who are due to choose their options at 14 or 16 or have recently done so. The department did not say who ran the poll or when.

In the same poll, 46% said they were concerned that advances in artificial intelligence could reduce job opportunities. More than half (52%) worry about having the right skills for future jobs. Nearly half (44%) said their peers look down on vocational and technical routes.

For pupils and parents, the department set out the dates for digital courses. V Levels, new vocational courses for 16 to 19-year-olds that are equal to one A level, start in September 2027. Digital is one of the first three subjects. Foundation Certificates in Digital also start then, for students who need to pass English and maths first.

The department announced in September that Vocational GCSEs could include Digital Tech. Some will be available from as early as September 2029. A consultation on them will be published before the end of the year, the department said. It has not said what a Digital Tech course would teach or whether it will be among the first subjects.

Xi Zhao, Xinyue Jiao and Zhen Xu wrote the paper. It uses 3,041 answers to 50 open questions from computer science courses at the Universidad Francisco de Vitoria in Spain in 2025 to 2026. A teacher marked each answer out of 10. So did three large language models (LLMs), the AI models behind chatbots: DeepSeek, Qwen and Gemini.

The models' scores agreed moderately to strongly with the teacher's, the full text says. Their written feedback differed more. Teachers flagged a misconception, an error in understanding, in 15.0% of comments and the models in 4.8% to 7.4%. For answers the teacher scored 0 to 2, the figures were 31.5% and 5% to 7%.

For lecturers and teachers who use AI to mark, the authors suggest it may suit procedural topics such as software testing. They say topics that need judgement should still be marked by people. In England, Department for Education guidance already says "Teachers, leaders and staff must use their professional judgement when using these tools."

The paper's header says it appears in the proceedings of the 34th International Conference on Computers in Education. The authors say the data comes from one institution and one software engineering course. They did not test newer models, so it is not known whether those give better feedback.

The Education Endowment Foundation (EEF) commissioned the study and the National Foundation for Educational Research ran it. Tes reported that it involved 464 primary teachers in 108 schools. One group was encouraged to use the tool, called Aila. The other was told not to use it but could use other AI tools.

Teachers given Aila spent an average of two and a half hours a week on planning and resources, Tes said. The other group spent three hours and 19 minutes. Oak said in a blog post that this saved an average of 49 minutes a week. EdTech Innovation Hub reported that independent assessors found the Aila group's materials were no worse in quality.

Use was low. The same report said 81% of the Aila group tried it at least once, but it was used for about 7% of lessons in an average week. Fewer than one in five teachers told researchers it had meaningfully influenced their teaching and planning, according to Tes. Many preferred easier AI tools.

For teachers, the main saving went to those early in their careers or less sure of their subject, Tes said. The EEF cautioned that saved time does not automatically mean better teaching, EdTech Innovation Hub reported. It said teachers must still check AI materials for accuracy.

Schools Week reported on 5 November that the Publishers' Association has written to Oak asking it to suspend Aila until it can show the tool works without members' materials. Oak said the association had not identified any specific materials where copyright was infringed, according to the report.

The trial data dates from autumn 2025, so it does not show how the tool performs now. Tes and EdTech Innovation Hub give no pupil results. It is not known whether Oak will suspend Aila.

The company said in its results post that more than 30,000 people took the survey over seven weeks. The Register reported 30,903 responses from 169 countries. Stack Overflow said 52% of people are still learning new coding skills. When they learn something new, 79% want to build AI skills.

Of the 14,530 people who said how they learned to code in the past year, 58.9% used technical documentation. These are the guides that the makers of a language or tool write about it. The results page shows AI code generation tools second at 52.6%. Online courses or certification were picked by 29.3%, school (university or college) by 14.2% and coding bootcamps by 6.3%.

For learners, the figures show that people who already code most often learn from documentation and AI tools. Fewer than a third used an online course or certificate. The Register reported that about 7% of respondents trust AI output for important work decisions. It said 48% trust it when they can easily check it.

People could pick more than one answer. The results page does not split the answers by experience. So it does not show how complete beginners learned or how well each method worked. Stack Overflow said the full data set will be published on GitHub in a few weeks.

Ofqual's announcement says that is down from 29% and 34% in the two years before. Of schools hit, 66% recovered immediately, up from 55%. Critical damage fell to 7%. Last year's poll put critical damage at 10%. It also showed immediate recovery at 63% in 2023 to 2024, so this year's figure is three points above that level.

The poll asked for the first time who is mainly responsible for cyber security. Of teachers who answered, 46% said the IT team and 40% said all staff. Only 9% said senior leaders. Department for Education standards make a senior leader the person accountable for this work. IT support carries it out.

Teacher Tapp ran the poll for Ofqual on 13 July 2026. It counts what 3,775 secondary teachers said. School figures use one answer per school, from up to 2,162 schools. Ofqual's announcement mentions colleges, but its sample is described as secondary teachers. Infosecurity Magazine reported the same headline figures.

For teachers and students, Ofqual's concern is that coursework or marks can be lost in an attack. It asks senior leaders to arrange a review with their IT lead and to use the department's cyber response plan template. It said 55% of secondary schools surveyed have already taken steps such as backups or risk assessments.

The announcement does not say how many schools lost coursework or marks. It does not say whether the 7% is a share of all schools or only of those hit. Ofqual offers the full poll only by email.

Teacher Tapp asks more than 10,000 teachers questions each day through an app. In its weekly results, 82% of secondary teachers and 68% of primary teachers said AI literacy should be covered in most or some subjects. Only 3% of secondary teachers and 5% of primary teachers said it should not be taught.

Practice is different. Only 20% of secondary teachers and 13% of primary teachers said their school covers AI literacy in most or some subjects. In secondary schools, 41% said it is not taught at all. In primary schools the figure was 56%.

Teachers' own use of AI is high. In results published on 13 October, 80% said they had used AI for schoolwork in the past week or month. That was up from 74% a year earlier. In the September results, 27% of classroom teachers who used AI tools said they did not save time, up 11 points since April 2025.

Teachers have asked for clearer rules. Tes reported on 9 September that a NASUWT official told the Commons Education Committee that schools want "much clearer rules of the road" on AI.

For pupils, some AI teaching is now required. Revised relationships, sex and health education guidance took effect on 1 September 2026. It asks schools to teach about deepfakes and the risks of AI chatbots. Its references to AI are about online safety and the law, not how AI tools work or when to use them in schoolwork.

Teacher Tapp's panel is not a random sample. Its post did not give the number who answered. It is not known whether the new national curriculum will include AI literacy.

The report was written by Ipsos and Perspective Economics for the Department for Science, Innovation and Technology. Core cyber adverts asking for less than two years' experience fell by 53% over those three years. Adverts asking for six to nine years fell by 23%. More than 70% of adverts did not state any experience level, so the report calls the trend indicative.

Graduate numbers kept growing. UK universities produced 7,950 cyber security graduates in 2023/24, up 14% on the year before. Starts on cyber security apprenticeships in England fell by 15% in 2024/25, from 590 to 500.

In interviews, employers said AI and automation were taking over routine monitoring work in security operations centres. Those jobs have long been a first step into the field. The 51 interviews were not a representative sample. In the report's survey, 70% of cyber firms said staff used AI in their daily work, up from 53% a year earlier.

For learners aiming at cyber security, the report says 12% of core cyber adverts were open to people with GCSE, A level or foundation-level education. Among cyber firms that had recruited since 2024, 65% said they had hired through routes that did not need a degree. Employers also said they received many junior applications from people without the skills the job needed.

The government's Digital and Technologies Jobs Plan, published in October, drew on the same report. The report says part of the fall may reflect firms changing how they recruit, such as using agencies. How much of it is caused by AI is not known.

The university said the study analysed survey answers from 3,721 secondary pupils and 352 secondary teachers in and around Bristol in 2025. The Nuffield Foundation funded it. Pupils told researchers they used phones in toilets, under tables and through earbuds or smartwatches.

Staff described the same pattern. The university said 61.3% of staff saw pupils go to the toilet to use phones. Nearly two-thirds (64.7%) enforced phone rules at least weekly and 29.3% did so daily or at least once per lesson.

Stricter rules may also meet resistance. The study found 93.5% of pupils wanted to keep "off and in bags" or to have fewer or no restrictions. Pupils who saw themselves as struggling most at school were the least keen on stricter rules, the researchers said.

The findings matter for schools because the rules have changed since the survey. Department for Education guidance on phones became statutory on 29 June 2026 and schools should follow it from 1 September. The university said the government has left schools to decide how to apply it. Tes reported that in Teacher Tapp polling in September, 52% of secondary academies and 54% of council-run schools had an "out of sight all day" policy.

The survey covers one area of England and relies on what pupils and staff reported. It was carried out before the guidance took effect. The university's announcement does not say whether the work has been peer reviewed. It is not yet known whether schools can enforce "off and in bags" more successfully under the new rules.

The report comes from the AI and digital education working group of the department's Science Advisory Council. It is independent advice, not government policy. Its authors include Professor Dame Athene Donald, who chairs the council. The department's chief scientific adviser, Professor Michael Thomas, is another.

The group says AI tools change faster than randomised controlled trials can test them. A trial's results may describe a version of a product that no longer exists. It cites a Stanford review of about 800 papers on AI in education, which found only 20 that met its standard for showing cause and effect.

Instead, the group wants evidence built up from many sources and updated as it arrives. This would include reports from teachers. It says caution should come first where there is a credible risk of serious harm to children. For lower-risk uses, such as tools that help teachers with admin, it favours learning as use grows.

Two principles affect schools most directly. One warns against buying decisions that create a dependence on a single supplier before the evidence is in. Another suggests AI suppliers could be required to give schools and independent evaluators usage data as a condition of public sector contracts.

For teachers, the report does not change any rule. It says the principles sit alongside existing department guidance. Inside government, they are meant as a shared reference for decisions about AI in education. For suppliers, they set out the questions about evidence that developers should be ready to answer.

The department has not said whether it will adopt the usage data proposal. Its safety certification plans, published on 10 November, would check AI products for safety but not for whether they help pupils learn.

The EdTech Testbed Programme guidance says the work will run for four years in more than 1,000 settings in England. A group led by Ipsos will run up to 100 rapid evaluations. The most promising tools will then go to larger studies.

The Education Endowment Foundation will spend £7m on up to 15 randomised controlled trials or other rigorous evaluations. A randomised trial compares pupils who use a tool with similar pupils who do not.

A separate report, published the same day, explains why the evidence matters.

The Chartered College of Teaching ran a pilot for the department to review the evidence behind EdTech products. The pilot was small. It received 96 expressions of interest and chose 13 products. Five of those used generative AI and had an initial review only.

The report said relatively few portfolios gave a clear account of whether a product works. It said AI products were being developed faster than strong evidence of their impact. It recommended considering changes to the review criteria for these products.

For teachers and school leaders, the department says the aim is evidence to support buying decisions. Schools and colleges will be recruited in phases, starting in the 2026 to 2027 school year. Those that registered for the earlier pilot do not need to register again at this stage.

The guidance does not name the products that will be tested or give a date for the first results. It also does not say whether the Evidence Board review will continue.

The Department for Education said the courses will be the same size as GCSEs and have nationally set content. Ofqual will regulate assessment and grading.

They will eventually replace Technical Awards, the current vocational courses for this age group. The department said these vary by provider and have no nationally set content. Other possible subjects include Creative Media Production and Building and Construction.

The department also said new subjects will come from a review of existing GCSEs and the new courses, to avoid unnecessary overlap. It did not say how Digital Tech would sit beside GCSE Computer Science.

LBC reported that Pepe Di'Iasio of the Association of School and College Leaders questioned the name and said the change alone will not remove barriers to take-up. Daniel Kebede of the National Education Union said success depends on sustained investment in schools and teachers.

For pupils and parents, nothing changes yet. Georgia Gould, answering for the department in a written answer on 3 November, said new technical pathways will at first use existing Technical Awards. She said the first Vocational GCSEs could be introduced in some subjects from September 2029. The department said English, maths and science will remain a common entitlement.

The department has not confirmed which subjects will come first. It has not published content for Digital Tech or said when Technical Awards will end. LBC reported that a consultation will follow later this year.

The first is a staff paper called "Friend or foe?", which reviews existing studies. It says AI can help students complete tasks but may not build knowledge. It says tools designed for learning generally do better than general-purpose tools.

The paper states that its views are not necessarily the Commission's own. It adds that most of the evidence comes from outside the EU.

The paper cites Eurostat figures showing that four in ten people aged 16 to 24 in the EU used generative AI for formal education in 2025. Use ranged from 59.4% in Spain to 17.6% in Germany. It also cites PISA 2025: 15-year-olds across the EU who often use AI to summarise texts scored 18 points lower in science than those who do not do so often. That figure is an association and does not prove that AI caused the gap.

The second is a Eurydice report covering 38 education systems, which do not include the UK. It says nearly two thirds refer to AI in top-level policy documents, binding or not. Around a third include AI skills in the lower secondary curriculum.

For teachers, both reports point to a training gap. The staff paper says 29.7% of teachers in the EU systems surveyed took part in training on AI for teaching in the 12 months before the TALIS 2024 survey. Eurydice says only seven systems include safe and ethical AI use in the skills expected of new teachers. The staff paper also says AI cannot replace a parent's everyday support.

The Commission said the findings will feed into a coming Education Package on basic skills, teachers and digital infrastructure. Its announcement gives no date. The staff paper says the long-term effect of generative AI on skills is still unclear.

The authors are the former head teacher Sir Anthony Seldon and the venture capitalist Tim Bunting. BBC News reported that their 10-point plan wants schools to write their own AI strategies by September next year. Ofsted would then assess those strategies during inspections. Seldon co-founded AI in Education, a group set up in 2023 to help schools and colleges use the technology safely, the BBC said.

The report includes a Public First survey of 4,021 people aged 11 to 24 in England, carried out in August. Tes reported that only 9% believe schools and universities are preparing them very well for a world with AI. That figure counts only the top answer. The reports do not give the other answers.

LBC reported that 87% of those surveyed have used an AI tool. It said 47% would trust AI over a person to check a fact. The report warns that some families and pupils may stop seeing school as a "legitimate use of their time and effort".

Other proposals include a rapid response unit on AI inside the Department for Education and a network of 100 "AI champion schools and colleges". The plan also proposes a "two-lane" assessment system. It would combine exams with continuous evaluation of pupils using AI.

For teachers and parents, nothing changes yet. These are recommendations and not government policy. A department spokesperson told the BBC that ministers take AI "incredibly seriously" and pointed to its curriculum changes, its AI tutoring tools and new vocational routes.

The report is a call for action by its authors. The reports do not say whether the department will adopt the framework or the September 2027 date.

The review's summary groups the risks into three areas. Educational risks include less reliable assessment and a weaker bond between teacher and student. Societal risks include harm to the Estonian language and values passed on by an outside provider.

Technology risks include data breaches, rising costs and service failures. The summary says these are scenarios and the harms may not happen.

AI Leap's announcement calls the review independent, but the foundation that runs AI Leap commissioned it. The authors include staff from the Estonian firm Cybernetica and the European Commission's AI Office. The full report names OpenAI and Google as technology partners. It says researchers from the University of Tartu, Stanford University and OpenAI are studying the programme's effects.

The report says only 50% of students had activated their accounts by 31 March 2026, slower than expected. In a survey at the end of the first school year, more than 90% of participating teachers said they used AI. They estimated that it saved about two hours a week.

The programme is now growing. This school year it aims to reach about 18,000 vocational students and up to 13,000 lower secondary teachers. Lower secondary pupils will not get the student app, because the programme first wants to learn which uses help at that age.

For teachers elsewhere, the review recommends judging such tools by whether they support learning. It says assessment and teacher training must change alongside the technology.

The first findings from the study of learning are expected by the end of 2026. AI Leap's announcement gives no timetable for acting on the recommendations.

V Levels are vocational courses for students after GCSEs. They will sit alongside A levels and T Levels in England.

Ofqual said in its announcement that 40% of the marks will come from timed exams and 60% from other assessment. A V Level is about the same size as an A level. Grades run from A* to E, then P for a pass.

The consultation outcome shows Ofqual first proposed on-screen exams only for this subject. It then added accounting and finance. Exam boards that offer both paper and screen versions will not need separate specifications. Ofqual's summary says respondents raised college readiness, digital inequality and the effect on learners with special educational needs. FE Week reported that Ofqual went ahead despite those concerns.

The course content, published by the Department for Education in July, covers AI-enabled systems and structured coding. It says students are not required to use a specific programming language. Their digital solutions may be scripted, automated or low-code.

For students choosing courses for September 2027, the rules set the shape of the assessment. For colleges, Ofqual's decision permits on-screen exams but does not require them. Ofqual's summary says most respondents favoured a gradual, phased rollout.

Ofqual still has to consult on the draft rules for on-screen exams. It will decide on future V Levels subject by subject. It is not yet known whether exam boards will offer this course on screen.

INTEF is part of Spain's education ministry. Its announcement, in Spanish, says the update responds to how fast the technology is changing. The guide is organised around three groups: pupils, teachers and schools. INTEF officially launched it in a live webinar on 5 October.

The new version puts more weight on AI literacy for pupils. It also covers privacy, fairness, academic honesty and the effect of AI on the environment and on pupils' wellbeing. The English edition adds three practical tools. They are sample learning activities for different stages of schooling, a template for pupils to declare how they used AI in a task and a questionnaire that helps pupils reflect on their use of AI.

The guide says teachers should set clear conditions for using AI in each task and ask pupils for a declaration. It says this builds trust and reduces misuse.

The guide also notes that under Spanish data protection law, children cannot consent on their own to the use of their personal data until they are 14. On marking, it says the EU AI Act treats AI that assesses learning as high risk. Such use is allowed but brings extra duties for whoever uses the system.

For teachers in Spain, the guide is advice. It does not change the curriculum or the law. The education ministry had said in a June press release that the update would come at the start of the 2026/27 school year. INTEF has not said how many schools use the guide or whether regional governments will adopt it.

PISA tests 15-year-olds in 91 countries and economies. The OECD said almost two-thirds of students in OECD countries reached the relevant level in computational problem solving. Only half reached it while also showing basic skills in science, reading and mathematics. The Scottish Government's report defines the skill as using tools such as programming and computer modelling to solve unfamiliar problems.

The same report says the UK scored 523 points against an OECD average of 500. England scored 525, Scotland 515, Northern Ireland 513 and Wales 501. Japan had the highest score among the OECD countries listed, at 557. In England, boys scored 537 and girls 512.

PISA also asked about AI. The OECD's UK note says 31% of UK students used AI chatbots to help them learn at least once a week, against an OECD average of 46%. The OECD said students who did not use AI to draft written work outperformed those who did. These are patterns in test and questionnaire data, not a trial. The Royal Society said the findings show an urgent need to understand AI's uses in education.

For computing teachers, this is the first PISA measure of these skills, so it shows no trend over time. It places no new requirement on schools. The OECD said UK exclusions from the test were above its tolerance, but that the data were fit for reporting. The Department for Education's report says England's sample may not be entirely representative.

The OECD's full results on this test are due in May 2027. Its analysis of how students direct their own learning has not yet been published.

He spoke to the House of Commons education committee's inquiry into AI and education technology. The transcript records him saying that AQA is not talking about using AI to mark high-stakes exams. AQA has piloted a system for two years. It watches markers as they work and alerts a senior examiner to problems such as inconsistent marking.

Hughes said AQA had 96 malpractice cases involving AI this year, which was 2.4% of its malpractice cases. Ofqual proposed in December 2025 that each exam board could offer up to two on-screen courses, but not in subjects with over 100,000 entries. Hughes asked for a wider start. Andrew Freeman of Pearson also asked Ofqual to go faster.

Sir Ian Bauckham, Ofqual's chief regulator, told the committee that the low number of detected AI cases must be treated with heavy caution. He said teachers tell him they know pupils use AI but cannot prove it.

His approach to extended writing coursework is to "avoid, adapt and deter". He said Ofqual has set up a new team on qualifications integrity. He also said AI will not be allowed to mark complex work unaided at any time soon.

For pupils and teachers, the hearing changed no rule. Bauckham said Ofqual is giving the Department for Education evidence on the risks of some extended writing coursework as it reforms qualifications. That could affect future courses, but no decision was announced.

FE Week reported on 10 September that a decision on on-screen GCSEs and A levels was expected later this year. Ofqual's consultation page still says it is analysing responses. The committee is still taking evidence and has not yet published a report.

K-12 Dive reported that the block covers all pupils for the 2026-27 school year. Before it, pupils aged 13 or over could use approved AI tools once they had finished lessons on digital citizenship. A district spokesperson said the block will stay while the district reviews how AI is used in teaching and what safeguards it needs.

LAist reported that web-filtering software now blocks any site or app classed as an AI tool on district Chromebooks. A similar tool covers tablets. LAist said the change affects an estimated 378,000 pupils.

The decision surprised the board. Education Week reported that senior administrators imposed the pause on their own authority. Board members Kelly Gonez and Nick Melvoin expressed surprise, though neither objected to it. Melvoin said the step had not come from the board. On the same day, New York City announced a pause on generative AI for its younger pupils.

For families in Los Angeles, the block applies to district laptops and tablets. LAist said the district has no rule yet on pupils' own devices. The filter still applies if a pupil signs in to a district account on a personal device. The district's interim chief information officer told LAist that the software also records how long pupils use district devices.

The district told K-12 Dive that a board committee will recommend rules for future pupil use of AI by the end of the 2026-27 school year. It is not known whether the block will be lifted before then or which tools might return.

The mayor's office said the moratorium runs from 2-K, the city's youngest year group, to grade 8, the last year before high school. It affects nearly 600,000 pupils, about two-thirds of enrolment.

It covers all software with student-facing generative AI. Companion chatbots are banned in every grade. The city called it the most expansive policy of its kind in the United States.

High school students will have two 45-minute AI lessons a year. Up to 50,000 of them will join five pilots under direct supervision from a trained teacher. amNewYork reported that the pilots leave out general chatbots such as ChatGPT. It also reported that the education department will turn off AI in nearly 40 programmes it had allowed. A product will be removed if its AI cannot be turned off.

The city said exceptions apply to assistive technology for pupils with disabilities and for multilingual learners. They also apply to pupils on career readiness programmes such as computer science. Teachers can still use AI for planning and operational tasks. CBS New York reported that they cannot use it to grade work. It quoted the United Federation of Teachers, a union, as saying the policy "leaves many questions unanswered".

The rules apply only to New York City's public schools. The city also recommended screen limits of 30 minutes a day in grades 3 to 5 and 45 minutes in grades 6 to 8. In November, Mayor Zohran Mamdani criticised Google over its plan to withhold Search from schools that ban AI.

A new Technology in Schools Coalition will assess the moratorium during the school year and publish recommendations. The announcement does not give a date for that report. It also does not say what will happen after the 2026 to 2027 school year.