Issue 10
Viewpoint
child safety · United States
Common Sense Media: ChatGPT for Teens poses an 'unacceptable risk' as safeguards fall short
A report published on 7 October by the Common Sense Media Youth AI Safety Institute rated OpenAI's teen version of ChatGPT an unacceptable risk. According to the report, parental protections fail: in more than 4,000 test messages, parents received no notification during suicide-related conversations.
Who is affected and how
- parent
- Parents should not rely on parental notifications, since the test found none were sent even on suicide-related topics; a child's ChatGPT use should be followed through direct conversation.
- school leader
- According to the report, the institute does not recommend use under 18, so school guidance and the question of banning or allowing it should be reconsidered.
- student
- Study mode and 'teen mode' do not guarantee safety; on sensitive topics, turn to a person.
school practice · United States
US universities turn away from AI detectors amid 'homework scams'
US universities reported in early October that detection tools have not worked, so they are using other methods to fight AI-assisted cheating. The topic is sharpening the debate over the authenticity…
teacherAn AI detector's result cannot be the only evidence; assessment should shift toward oral and in-person formats.university studentBecause the tools used to raise suspicion are unreliable, it is worth keeping evidence of your writing process.education policymakerPolicies built on detectors should be reviewed.AI companion · United StatesKentucky sues Character.AI, alleging chatbots encouraged self-harm and starvation
Kentucky's attorney general filed a lawsuit against Character.AI on 8 October, alleging its chatbots encouraged users, including minors, to self-harm, starve themselves and take their own lives…
parentAI companion apps can also produce self-harm content, so it is worth checking how a child uses such apps.education policymakerA state-level lawsuit signals that companion chatbots are being treated as a liability issue.regulation · United StatesFrederick County school board in Maryland bans student AI use from pre-K through 12th grade
The Frederick County (Maryland) school board voted to ban students' use of AI from pre-K through 12th grade. Another report says a district switched to a blanket ban within weeks of rolling out…
school leaderPositions on an AI tool can change quickly between rollout and ban, so a regulatory framework is needed before expansion.teacherUnder a ban, AI-based assignments and materials cannot be used with students.parentParents can have a say in AI rules at the public hearings.regulation · United StatesNew York City and Los Angeles school systems rein in student AI use
According to Education Week, in September the two largest US districts, New York City and Los Angeles, curbed students' use of AI. In Los Angeles the restriction is under debate, while in New York…
school leaderThe example of the two largest districts puts pressure on other systems, and whether the rules can be enforced is an open question.studentStudents use AI despite the ban, so it is worth knowing the rules.Hungary
school practiceAI may rewrite university assessment as the age of take-home assignments could endteacherCourse assessment methods need to be redesigned, because the traditional take-home assignment increasingly fails to demonstrate individual knowledge.university studentAssessment may shift from take-home assignments toward in-person, verified independent work.regulationLannert Judit: Hungary's education strategy to be ready by year-end, AI risks debatededucation policymakerThe strategy promised for year-end will define the role of AI, and positions on risks may shape the direction of regulation.school leaderBased on the forthcoming strategy, it is worth thinking ahead about institutional AI rules.studentAdvice on the labour market and learning is disputed, but the emphasis is on deep study and reading.surveySurvey: two-thirds of young Hungarians use AI for emotional supportparentA child may also seek emotional support from a chatbot, so it is worth asking openly what they use it for.teacherSchool mental health work must take into account that students may confide their problems to AI.surveyIdea Institute: school is the biggest source of stress for 14–19-year-olds, many ask AI for adviceschool leaderSchool as the main source of stress raises the question of reviewing workload and the support system.parentBecause of shame, a child turns to AI more easily, so a low-threshold, non-judgmental conversation is needed.school practiceConference for teachers to discuss how AI could change Hungarian classesteacherThe structure of lessons and the teacher's role may change under the influence of AI, and training sessions offer a chance to get to know the tools in advance.school practiceWill theses survive the age of AI? Hungarian debate over the 'AI pencil case'university studentThe form of the final thesis and the rules on AI use may change, so clarify with your supervisor in advance what is permitted.education policymakerAn institution-level decision is needed on the role of the thesis and on alternative assessment.school practiceEdTech Summit Hungary: who is responsible when AI and the teacher grade differently?teacherEven with AI assessment, the teacher remains responsible for the final grade.edtech companyAssessment tools must be transparent and open to override by Hungarian teachers.surveyHungary leads the Visegrád countries in AI use and ranks 20th globallyeducation policymakerWith high usage, improving AI awareness is the next task.teacherAI is spreading among students and instructors, but conscious use does not come automatically.school practiceÓbuda University announces AI-native strategy to renew teaching and examinationsuniversity studentAt Óbuda University, the strategy may affect how students learn and how exams are held.teacherInstructors will need to adapt course material and assessment to AI-native operation.toolGoogle and Codecool offer over 3,000 free AI training places to Hungarian professionals and studentsuniversity studentStudents can apply for a free place and a certificate that can be listed on a CV.edtech companyA training model built on Google and run with a local partner may compete with the domestic offering.school practiceSopron's Széchenyi Grammar School regulates rather than bans AI use in schoolschool leaderA Hungarian grammar school provides a concrete example of regulation without a ban.teacherThe policy gives uniform guidelines for designing assignments.studentRegulated use means students need to know which AI use is permitted and which is prohibited.school practiceAI also helps in BME maths and physics preparatory coursestudentThe course can provide admission points and possibly later credits, and is available outside Budapest as well.parentThe preparatory course is one alternative for preparing for university admission.school practiceAI workshops for instructors: University of Szeged autumn series and chamber event in SzegedteacherIn the series, University of Szeged instructors can learn about practical questions, such as how students and instructors themselves prepare for classes.school practiceCzech inter-university AI programme prg.ai Minor continues as interest growsuniversity studentAn inter-university AI minor is available to students.caseTata school principal: AI can wean children off thinking altogetherschool leaderStudent views can also be gathered for school AI policy, as was done in Tata.teacherPreserving independent thinking is a central consideration in designing assignments.toolFree AI course with adult-education micro-credential launched at Dennis Gabor Universityuniversity studentA free, short course on using AI at work, with a micro-credential.fraudHungarian Competition Authority tours schools with renewed programme on manipulative ads, scams and algorithmsteacherSchools can sign up for the programme, which provides material for digital literacy lessons.studentThey learn to recognise advertisements, scams and algorithms.school practiceSemmelweis University takes part in Zhejiang University anniversary symposium on AI in dental educationuniversity studentInternational cooperation on AI in education is being built into dental training.
More key stories of the week
research · global
World Bank: AI raises homework scores but lowers exam results
Studies show that AI use improves homework scores but reduces exam results. International evidence cited by the World Bank indicates that unrestricted use of generative AI improves submitted work…
teacherHomework scores do not reflect knowledge; exams and in-class assessments remain necessary.studentA good homework score earned with AI is no guarantee of success in an exam.education policymakerAn argument against unrestricted AI use in the regulatory debate.survey · United StatesSurvey: 56% of US children have an AI app, with AI companions holding their attention 93 minutes a day
A report published on October 6 says 56% of US children now have an AI app, and AI companions hold their attention for 93 minutes a day. The figures show how widespread companion chatbots have become…
parent93 minutes a day is a significant amount of time; it is worth checking children's phones for AI companion apps.school leaderA substantial share of students use AI apps, so digital awareness should be part of the curriculum.tool · United StatesChatGPT for Teens gets a college planner and new study tools
OpenAI has announced that ChatGPT for Teens is getting new study tools and a College Planner feature. The planner will initially help US students in grades 10–12 applying to four-year colleges to…
studentTracking admission deadlines and aid could move to one place, but for now the feature is only for US users.parentBefore leaving college-application planning to AI, check what data the child enters.case · United StatesOpenAI says teens use ChatGPT little; watchdog groups dispute the claim
OpenAI says teenagers use ChatGPT for less than 15 minutes a day, a claim watchdog groups criticized. The company's response came on 7 October to the Common Sense report, which says the product still…
parentThe company's argument based on average time does not refute the report's safety findings, so the settings should be checked regardless of usage time.education policymakerUsage data comes from the manufacturer while the risk is stated by an independent evaluator; in the regulatory debate the two should be treated separately.regulation · United StatesCalifornia signs Adam's Law on companion chatbots as xAI minors case heads toward consolidation
California signed Adam's Law, which requires companion chatbots to have crisis protocols, parental controls and protections for children. Separately, a request has been made to consolidate x.AI…
edtech companyChatbots aimed at children must include mandatory safeguards, such as crisis protocols and parental controls.education policymakerThe California model could serve as a template for regulation in other states and countries.research · United KingdomEEF: AI tool reduced primary teachers' lesson planning time
The Education Endowment Foundation (EEF) says an AI tool reduced primary teachers' lesson planning time. The summary does not state the size of the reduction.
teacherAI-assisted lesson planning can free up time; the exact size of the effect can be checked in the full report.school leaderThe result may support the case for introducing AI tools.tool · globalGoogle extends Gemini student hub to K-12 classrooms of all ages as Gems move to Skills
Google is extending the Gemini student hub to K-12 classrooms of all ages. At the same time, Gems is automatically switching to Skills for education users, and according to Gemini help, Skills is for…
teacherPreviously assigned Gems are moving to Skills; it is worth checking what happens to assigned items.school leaderBecause of the age requirements, check that settings are appropriate for the product before rollout.edtech companyGoogle's platform changes affect Classroom integrations.research · United StatesAI tutoring trials: University of Maryland study raises more questions than answers
The University of Maryland ran a randomised controlled trial of its own AI tool; according to Inside Higher Ed, the results raised more questions than answers. Other trials, including Khanmigo's…
school leaderBefore introducing AI tutors, note that the evidence so far is mixed and depends on design.edtech companyThe impact of tutors needs to be demonstrated in controlled trials.teacherThe benefit of an AI tutor depends on how it is designed.regulation · AsiaMalaysia says AI chatbots for children must be safe by design
Malaysia's government says AI chatbots aimed at children must be designed to be safe from the outset, and child protection will be built into the proposed AI Governance Bill. Digital Minister Gobind…
edtech companyCompanies making chatbots for children should expect design-level safety requirements.education policymakerThe effectiveness of the regulation is disputed, as one article asks.school practice · United StatesUS school AI rules: Belmont dispute, Oklahoma limits, South Bend usage rates
Students and parents petitioned for clearer rules over the AI policy at Belmont High School. Oklahoma schools are setting limits on AI, and school officials in South Bend say more than half of…
school leaderClear rules, consequences and AI education are needed, as the Belmont petition requests.parentParents can have a say in shaping school policies.teacherUse is widespread, so clear classroom rules are needed.child safety · United KingdomUK data show rise in AI-generated child sexual abuse images
According to UK data, AI-generated child sexual abuse images are rising sharply in 2026. A Guardian editorial says the extent of image-based abuse in schools needs to be examined.
parentParents need to know about the risk and the ways to protect their children.school leaderImage-based abuse affects school incident-handling protocols.teacherTeachers cannot deal with the phenomenon on their own.regulation · European UnionNetherlands to introduce full smartphone ban in schools from January 2027
The Dutch government is introducing a full smartphone ban in schools, with the new rule taking effect from January 2027. The news was published on October 6.
education policymakerAnother country has joined the list of European countries with school phone bans.school leaderSchools need to prepare for enforcing the ban.In brief
researchMIT report: AI is eroding office hours, study groups and trust between faculty and studentsteacherThe value of personal contact with students and of office hours needs to be deliberately protected.university studentAI can replace human learning communities, which also affects learning itself.fraudAI glasses solve Suneung exam question during South Korean parliamentary audit, raising cheating concernseducation policymakerWearable AI devices threaten the security of national exams.studentWearable AI devices may be added to the list of items banned in exams.child safetyFBI and French authorities seize websites selling deepfake child sexual abuse materialparentChildren's photos and videos uploaded to social media can be targets of abuse.school leaderProtecting students' likenesses is also a consideration in school communications.regulationSpanish universities to make AI use mandatory in education, Hungarian report sayseducation policymakerMandatory use, rather than a ban, may also come up as an option.university studentAt Spanish universities, AI use may become part of the curriculum.caseAI is already eliminating university majors in China, Hungarian report saysuniversity studentAI also changes the risks involved in choosing a major.education policymakerCourse offerings need to be redesigned in light of AI's effects.toolEdtech funding: ByteXL raises $9 million, Arivihan in talks for Rs 96 crore, Ed.ai raises €5 millionedtech companyInvestors are backing AI coaching, admissions and assessment products.school leaderProviders are raising funds quickly, so procurement decisions should also weigh their longevity.caseUS charity plans to use AI to monitor its schools in Gazaedtech companyIn a war zone, AI-based monitoring raises data and ethical questions.regulationUAE: AI literacy curriculum, AI and mobile phone rules, official stance on AI in schoolsschool leaderThe KHDA rules may apply to every school in Dubai.teacherAccording to the official stance, AI does not replace the teacher.caseParents use AI to remove watermarks from school photos to save £25parentRemoving the watermark is a legal and ethical issue, and affects the photographer's service.school leaderSchool photographer contracts and data protection practices should be reviewed.toolGlobe brings Khan Academy AI learning to more Filipinos through GoLEARNteacherCan assign tasks and goals and track progress and learning gaps.studentThe Khanmigo AI tutor supports learning.toolGoogle opens free AI training for teachers in the UK and Ireland, and via Digital PromiseteacherFree training is available on using AI tools in education.edtech companyGoogle is expanding into educator training as well.child safetyIndia should put child safety first in AI education, WEF expert says as AI companions come under scrutinyeducation policymakerChild-protection requirements should be set out in advance when AI education is introduced.toolAmazon's new kids' tablets come with an AI story-creation featureparentThe kids' tablet offers AI creative features and screen-time controls, which are worth reviewing before buying.researchEDUCAUSE 2026: AI tutors, assessment that tracks reasoning, and the role of human connectioneducation policymakerAI and resource shortages top the higher-education IT priorities.teacherAssessment is shifting toward tracking students' thinking.regulationInter-Parliamentary Union calls for stronger parliamentary action to protect children's rights in the age of AIeducation policymakerThe appeal urges legislators to protect children from AI-related harms.researchBocconi study: ChatGPT and critical-thinking training help students in different waysteacherTeaching critical thinking separately and using AI have different effects, so it is worth planning them separately.school practiceScotland's first AI teacher warns pupils not to trust everything it tells themstudentThe AI teacher's claims need to be checked.teacherCritical use of AI can be part of teaching.regulationMalaysia aims for 30,000 AI graduates by 2030university studentState support for AI degree programmes is growing.education policymakerA specific output target provides a benchmark.researchResearch and Turnitin data show how students really use AI to writeteacherTeachers need to know students' actual AI writing habits in order to set assessment rules.school practiceSchools and teachers return to written and conversation-based assessmentteacherShowing the learning process and an oral conversation are alternatives to traditional take-home assignments.studentExplaining one's own thinking may become more important in assessment.regulationChina's Ministry of Education launches nationwide AI literacy training for teacherseducation policymakerA nationwide teacher AI training programme is an example of state-level preparation.regulationKarnataka plans to start AI training in government schools from Class 6studentClass 6 students in government schools can expect AI training.education policymakerA state-level example of building AI into the curriculum.toolTikTok and Meta expand AI parental controlsparentIt is worth switching on the new supervision settings and talking them through with the child.toolCommon Sense Privacy launches AI privacy assessment for edtech companiesedtech companyThe assessment can be used to demonstrate the privacy compliance of AI features.school leaderThe assessment may also be a factor in procurement decisions.caseWhy college applicants sound like AI even when they are not using itstudentOriginal writing can also look suspicious as AI-generated, so it is worth keeping traces of the writing process.teacherIn assessing admissions texts, style alone is not sufficient evidence.school practiceNew Chicago school uses AI to teach students, Hungarian reports sayparentThe AI-led school model differs from traditional teaching, and details require further inquiry.teacherAI-led instruction affects the teacher's role.case16-year-old who planned hike with AI stranded on dangerous cliffparentOutdoor plans children get from AI should also be checked against human sources and a map.studentA chatbot is not a safety adviser.regulationChatGPT text to carry invisible watermark in the European Union, Hungarian outlet reportsteacherThe watermark could help identify AI-generated text, but details are not known from the available excerpt.toolAnthropic offers eligible teachers one year of free Claude access if they register by 30 June 2027teacherEligible teachers get Claude Pro free for one year; they should review the data-handling terms before registering.school practiceTeachers in Atlantic Canada sound alarm over student AI useteacherCanadian teachers' concerns are similar to experiences elsewhere in the world.school practiceChatbot used to teach children to question AIteacherCritical AI use can be practised together with the chatbot.caseUnpatched server behind Victoria student data breachschool leaderKeeping servers patched is a basic requirement in school data management.