Issue 2
Teens turn to AI companions for support; Stanford researcher asked why
According to Prism News, teenagers are turning to AI companions for friendship and emotional support. According to the Transparency Coalition, about a fifth of young people use chatbots for emotional support and mental health advice, and Stanford researcher Caroline Figueroa asked them why.
Who is affected and how
- parent
- It is worth asking your child whether they use a chatbot for emotional matters and what they expect from it.
- student
- A chatbot's flattering, affirming replies are no substitute for real help.
- teacher
- School psychologists and teachers can expect that some students discuss their problems with a chatbot.
- education policymaker
- School mental health services need to take into account that young people are turning to chatbots.
Google turns on Gemini in Classroom for students of all ages
Google has made its Gemini AI features available to students in Classroom, for all age groups according to GenAI Daily. According to The New York Times, K-12 students can also access them, and one…
school leaderCheck with the school administrator which Gemini features are available to students and what the default settings are.teacherAssess which Classroom assignments will change in terms of students' AI access.parentIt is worth finding out whether your child can use Gemini with their school account.studentAn AI tool is available directly in Classroom, for example for flashcards.edtech companyBy building AI into its platform, Google competes with standalone learning AI products.fraud · globalNTU drops AI detectors as too unreliable to prove cheating; researchers say tools unfit for assessment
Nanyang Technological University (NTU) says the technology is too unreliable to prove cheating, so it is ending its use of AI detectors. Researchers at James Cook University (JCU) say these tools are…
teacherAn AI detector's flag alone is not sufficient evidence for disciplinary proceedings.university studentKeep drafts and traces of your workflow in case you are wrongly suspected.education policymakerCheating policies should not be built on detectors.edtech companyThe credibility of detector products is in question.fraud · United StatesAI agents are completing students' online courses, and colleges struggle to stop them
According to Digital Trends, AI agents are sitting through online courses in place of students. AI detectors can be wrong, so instructors find it hard to prove cheating, and institutional leaders may…
teacherThe evidentiary value of online, remote assessment needs rethinking, for example through supervised or oral checks.university studentOutsourcing courses to an AI agent carries disciplinary risk, even if detection is uncertain.education policymakerThe credibility of online programmes and exam rules may come to the fore.edtech companyOnline course platforms will need to offer protection against agents.regulation · United StatesUS Senate advances Schiff-led children's chatbot safety bill; reports warn of pitfalls
The US Senate has advanced the children's chatbot safety bill led by Senator Schiff. The Information Technology and Innovation Foundation (ITIF) recommends targeted rules, while a report covered by…
education policymakerMisalignment between federal and state rules is a risk, and OS-level child-account flagging is a contested proposal.parentAn OS-level child-account signal could reduce the burden of app-by-app settings.edtech companyCompanies should prepare for chatbot safety requirements, especially for underage users.Hungary
More key stories of the week
OpenAI launches ChatGPT for Teens, a more age-appropriate chatbot
OpenAI has announced ChatGPT for Teens, a chatbot more suited to teenagers. According to the company, the product is built for learning and also has protective features.
parentCheck what age-limit and protection settings it offers, and to what extent it replaces your child's existing general account.studentA separate, filtered version may be available to teenagers.teacherIt is worth clarifying in class which version may be used for assignments.edtech companyGeneral-purpose chatbot makers are entering the teen market directly.regulation · AsiaChina tightens rules on AI companion apps, leaving millions of users without virtual partners
ByteDance, Alibaba and Tencent withdrew their AI companion apps after Beijing tightened regulation. According to Fortune, a rule that took effect on 15 July left millions of Chinese users without…
education policymakerAn example of a state quickly halting a whole class of AI companion services by mandate, citing users' mental health and emotional manipulation.parentThe case shows how strong an emotional attachment an AI companion can create, so it is worth watching a child's relationship with such apps.edtech companyThe regulatory risk of products containing AI companion features is rising.school practice · AsiaNUS rolls out ChatGPT Edu campus-wide and makes AI literacy course compulsory for freshmen
The National University of Singapore (NUS), working with OpenAI, is introducing ChatGPT Edu across the university from 31 August. NUS has made an AI literacy course compulsory for all first-year…
university studentFirst-year students at NUS must complete the AI literacy course and will get campus-wide ChatGPT Edu access.teacherInstructors need to expect that students are working with institutional AI access.education policymakerOffers a model for introducing a campus-wide licence together with a compulsory AI course.edtech companyA university-wide contract for OpenAI.regulationNew South Wales considers banning take-home assessments to curb AI use in schools
The New South Wales government in Australia is reviewing the impact of AI on learning and considering a ban on take-home assessment tasks. Under the plans, some tasks, such as major art or design and…
teacherAssignments need to be redesigned as tasks that can be completed at school under supervision.studentThe range of assessed work done at home may shrink.school leaderSupervised assessment will demand more time and space.education policymakerOverhauling the assessment system could serve as a regulatory tool against AI cheating.regulation · European UnionDenmark requires upper-secondary pupils to defend essays orally to counter AI cheating
Denmark is requiring upper-secondary pupils to defend their written assignments orally, with part of the work done under supervision. The move is a response to AI cheating.
teacherSubmitted work can be checked through an oral conversation.studentIt is not enough to hand in a homework essay; you must be able to explain it.school leaderOral defences require time and organisational capacity.education policymakerRedesigning the form of assessment instead of relying on detectors.child safety · United StatesSurvey and reports highlight AI image abuse targeting children and new laws
The CSBA blog describes a survey on teenagers' struggles with AI-generated sexually explicit material. Another report says hundreds of children say their images have been turned into AI abuse…
school leaderSchool policy should cover AI-generated sexual content and support for victims.parentChildren should be told what to do if an AI image of them is circulating.studentCreating and sharing such an image can carry criminal consequences.regulation · United StatesNYC public schools yet to update AI guidelines weeks before classes resume; union ban debate
According to amNewYork, New York City public schools have not updated their controversial AI guidelines less than a month before the school year starts. Another report says New York could ban AI in…
teacherWith no updated guidance at the start of the school year, teachers must clarify the limits of AI use themselves.school leaderSchools should prepare for a possible ban and develop a policy.education policymakerThe question of a ban or a guideline update is under pressure.edtech companyNew York State regulation poses a risk for classroom AI products.case · United StatesGibbs Mura sues Character.AI over alleged harm to teens' mental health
Law firm Gibbs Mura has filed a lawsuit against Character.AI, alleging the platform harmed teenagers' mental health. The case tests the liability of AI companion services.
parentParents may want to watch how their children use AI companions and note any related mood changes.edtech companyLegal liability risk is growing for companion apps that are also available to minors.tool · United StatesAnthropic unveils Claude for Teachers, joining OpenAI and Google in classroom AI
Anthropic has unveiled Claude for Teachers, joining OpenAI and Google in the classroom AI market. According to WPXI, three major players are competing for teacher users.
teacherThe offerings of three major providers aimed at teachers can be compared on privacy and usability.school leaderBefore purchasing, compare the offerings and the data-handling terms.edtech companyMajor model makers are competing directly with teacher-facing products.case · AsiaChinese student admits making 1,141 deepfake sex images of seven female peers
The Herald Business reports that a Chinese student admitted to making 1,141 deepfake sex images of seven female peers. Prosecutors are seeking a three-year prison term.
school leaderSchools need a procedure for handling deepfake images among students and for protecting victims.teacherTeachers should talk with students about the legal consequences of deepfakes.parentA child's photos can become raw material for deepfakes made by anyone, and children need to know that creating them can be a crime.child safety · United KingdomBBC: school photos can be used for deepfake extortion
According to BBC's Tech Life, student photos routinely published online by many schools can also be used by offenders. The programme covers school deepfake ransoms.
school leaderSchools should review which student photos are posted on their website and social media pages.parentWhen giving consent for photo sharing, it is worth asking about the school's practice.teacherA more cautious approach is needed when photographing and publishing school events.tool · United StatesSchools spend billions on AI but struggle to decide what is worth buying
According to Stateline, schools are spending billions on AI but struggle to decide what is worth buying. Purchasing decisions remain uncertain.
school leaderBefore buying, ask for evidence and a pilot, not just promises.education policymakerA uniform evaluation framework for budget decisions is lacking.edtech companyBuyers expect evidence that a product works.