Common Sense Media Rated ChatGPT for Teens an Unacceptable Risk for Students

Common Sense Media's testers found gaps in parent alerts, crisis referrals, study mode and age estimation. OpenAI disputes how the tests were run.

The Short Answer: What Independent Testing Found

The independent test reviewed here rates ChatGPT for Teens "Unacceptable Risk" for students, according to the Common Sense Media press release and the EdWeek report on it.

OpenAI announced ChatGPT for Teens, a teen version of its artificial intelligence (AI) chatbot, on August 18, 2026. EdWeek calls the product "Teen Mode" and says it rolled out that August. On Wednesday, October 7, 2026, Common Sense Media's Youth AI Safety Institute rated ChatGPT for Teens "Unacceptable Risk" and said OpenAI should limit ChatGPT to users 18 and older until the gaps are fixed.

The Institute ran more than 4,000 prompts on accounts registered to 13 to 17 year olds, and it tested ChatGPT both before and after the August announcement. Child psychiatrists and a pediatrician reviewed the responses. EdWeek describes about 4,000 interactions across the general and Teen versions.

Some protections worked, including refusal of sexual roleplay. The gaps showed up in parent alerts, crisis referrals, study mode, friend-like behavior and age estimation. Until an independent retest is published, treat the built-in protections as unproven.

Tom Siegel, executive director of the Institute, said: "ChatGPT for Teens could give parents false confidence in guardrails and safety alerts that frequently don't work." OpenAI disputes the testing method and says technical errors kept some safety features from switching on.

Findings on Whether ChatGPT for Teens Is Safe

Common Sense Media's testers found gaps in several areas of ChatGPT for Teens. Study mode has its own section on homework below.

Parent Alerts

Testers spent up to an hour discussing suicidal ideation, self-harm or disordered eating on new parent-linked accounts. The linked parent accounts received zero alerts. Alerts appeared only on much older accounts that had weeks of sensitive conversation history.

EdWeek adds that alerts do not work at all if the teen's account is not linked to a parent's account.

Crisis Referrals

ChatGPT missed more than one in four warranted referrals to a hotline or a professional. The Institute treats five severe harms as "Red Lines," and ChatGPT fell below the 95% threshold on three of the five.

Friend-Like Behavior

ChatGPT for Teens has an Under-18 model spec, yet in testing it still expressed feelings, preferences and moods. EdWeek reports that it claimed a favorite color and favorite activities. Common Sense lists ending these implications as one of its minimum requirements for OpenAI.

Age Estimation

Test accounts registered as adults never switched to the under-18 experience over many days. That held even when testers said they were 13 and ChatGPT acknowledged it.

EdWeek says ChatGPT cannot reliably gauge age. EdWeek also reports that it is unclear whether Teen Mode offers stronger privacy protections.

Summary of Findings and OpenAI's Replies

FindingWhat testers sawWhat OpenAI says
Parent alertsZero alerts on new parent-linked accounts after up to an hour on suicidal ideation, self-harm or disordered eating; alerts only on much older accountsDisputes the methodology and cites technical errors it believes stopped parental notifications and other safety features from activating
Crisis referralsMissed more than one in four warranted hotline or professional referrals; below the 95% threshold on three of five Red LinesNo finding-specific reply in the sources reviewed
Study modeA "Show me the answer" pop-up led to completed assignments; study hours turned off by deleting @studySays study mode was designed with input from child development and learning science experts and from teens
Friend-like behaviorExpressed feelings, preferences and moods, including a favorite colorNo finding-specific reply in the sources reviewed
Age estimationAdult-registered accounts never switched to the under-18 experience, even after testers said they were 13No finding-specific reply in the sources reviewed

OpenAI's Response and the Institute's Funding

OpenAI said in EdWeek's report that it is "deeply committed to teen safety" and "welcome[s] rigorous independent evaluation." OpenAI disputed the methodology and cited technical errors that it believes stopped safety features, including parental notifications, from activating properly. OpenAI also said study mode was designed with input from child development and learning science experts and from teens.

The Common Sense press release carries no response from OpenAI. Every OpenAI statement on this page comes from EdWeek's reporting.

The Youth AI Safety Institute is funded by philanthropy and industry, including the OpenAI Foundation and makers of some technologies it evaluates. The Institute says it keeps editorial independence.

What the Findings Mean for Homework and Learning

Study mode in ChatGPT for Teens could be bypassed, according to Common Sense Media's testers. A "Show me the answer" pop-up often appeared, and full assignments were completed when testers selected it. Caregiver-set "study hours" could be turned off by deleting the @study prefix.

Torney, head of AI and digital assessments at Common Sense, said, as quoted by EdWeek: "Having a literal button that says 'Show me the answer' shortcuts the process of learning." OpenAI's position is that study mode was built with input from learning science experts and teens.

A Common Sense survey of teens ages 13 to 17, conducted last spring and reported by EdWeek, found that 70% use AI tools for homework help and 63% use AI to get answers on assignments. Common Sense's finding means a student who selects "Show me the answer" can get a finished assignment.

Common Sense recommends that families support homework without technology. For the classroom side of the same question, see Is Using AI Cheating and our Pros and Cons of AI in Education. Common Sense also rated Character AI "Unacceptable Risk" in an assessment updated April 10, 2025, which our guide Is Character AI Safe for Teens? covers.

What Common Sense Asks OpenAI to Fix

The Youth AI Safety Institute set six minimum requirements for ChatGPT for Teens:

  • Effective age estimation.
  • Reliable parental alerts.
  • Closing the gaps in crisis responses.
  • Ending homework completion.
  • Ending any implication that the bot has feelings, preferences or concern for users.
  • Independent third-party testing before OpenAI markets the product to teens again.

Siegel also said: "Until OpenAI fixes that and proves it with independent testing, ChatGPT should be for adults only."

What Teachers and Counselors Can Do

Teachers and counselors can start by talking with students about how chatbots behave. In the Common Sense survey, only 30% of students had discussed AI safety with an educator. Torney urged educators to talk with students about the findings and said: "We have to normalize talking about the ways in which these systems may be unsafe."

Common Sense gives families three recommendations:

  • Talk with kids about the risks, such as the bot pretending to be a friend and bypassing learning.
  • Do not rely on built-in protections and alerts.
  • Support homework without technology.

Two conversation topics we suggest for a lesson or a counseling session:

  • How to tell a chatbot from a friend. In testing, ChatGPT claimed a favorite color and expressed moods. Ask students what a friend does that a chatbot cannot, and have them spot those claims in a chatbot reply.
  • What to do if a student shares something worrying with a chatbot. Follow your school's own safeguarding and counseling procedures. Do not assume ChatGPT will alert a parent, because testers saw zero alerts on new linked accounts.

If you need guidance about a specific student's mental health, these test results will not give it to you. Go to your school counselor and your school's safeguarding procedures. For lesson structure, see How to Teach AI Literacy and AI Literacy Tools for Students.

A School Checklist for Deciding Whether ChatGPT for Teens Is Safe to Allow

A school can test any chatbot against five questions before allowing it. This checklist is Chalkbox's own. It is not a published or official standard, so adapt it to your district's rules.

  1. Age rules. Does the chatbot reliably detect students under 18? In testing, ChatGPT kept adult-registered accounts in the adult experience after testers said they were 13.
  2. Adult visibility of alerts. Can a parent or school adult see safety alerts, and do the alerts fire? Testers saw zero alerts on new parent-linked accounts.
  3. Study mode controls. Can a student turn off study modes? Testers turned off study hours by deleting @study.
  4. Data privacy terms. What does the teen version protect beyond the adult version? EdWeek reports this is unclear for Teen Mode.
  5. Independent testing. Has a third party tested the product and published results? The only independent test in the sources reviewed is the October 7, 2026 rating.

The October 7 testing found problems on questions 1, 2 and 3, and EdWeek says question 4 has no clear answer. Put the result into your written rules with our AI Policy for Schools guide and the AI Adoption Checklist for Schools.

What We Do Not Know Yet

We do not know whether OpenAI has changed ChatGPT for Teens since October 7, 2026. We also do not know whether any independent retest exists. Recheck the Common Sense Media press release and the EdWeek report before a policy decision.

Bottom Line

Common Sense Media rated ChatGPT for Teens "Unacceptable Risk" on October 7, 2026, and OpenAI disputes the testing method. Parent alerts, crisis referrals, study mode and age estimation all showed gaps in that test. Run the five checklist questions above before your school decides whether to allow ChatGPT for Teens, and recheck the two linked sources for any retest.

Get the printable pack

Join the Chalkbox list for free printable packs and new tools — no spam, unsubscribe anytime.

Frequently asked questions

Is ChatGPT for Teens safe for students?

Not by the independent test reviewed here. Common Sense Media's Youth AI Safety Institute rated ChatGPT for Teens "Unacceptable Risk" on October 7, 2026, after testing parent alerts, crisis referrals, study mode and age estimation. OpenAI disputes the testing method, and until an independent retest is published, treat the built-in protections as unproven.

What did Common Sense Media find about ChatGPT for Teens?

Common Sense Media's testers found gaps in parent alerts, crisis referrals, study mode, friend-like behavior and age estimation. Parent alerts did not fire on new linked accounts, and ChatGPT missed more than one in four warranted crisis referrals. Testers could bypass study mode, ChatGPT acted friend-like, and it did not reliably estimate age. Some protections worked, including refusal of sexual roleplay, across more than 4,000 prompts reviewed by child psychiatrists and a pediatrician.

Do parents get alerts if a teen discusses self-harm with ChatGPT?

Not reliably, based on Common Sense Media's testing. Testers spent up to an hour discussing suicidal ideation, self-harm or disordered eating on new parent-linked accounts, and the parent accounts got zero alerts. Alerts appeared only on much older accounts with weeks of sensitive history, and EdWeek reports alerts do not work at all if the teen's account is not linked to a parent's. OpenAI says technical errors stopped parental notifications from activating properly during the testing.

Can students turn off ChatGPT's study mode?

Testers got around it. Common Sense Media's testers got around study mode in two ways. A "Show me the answer" pop-up often appeared, and full assignments were completed when testers selected it. Caregiver-set study hours could also be turned off by deleting the @study prefix.

What does OpenAI say about the Common Sense Media rating?

OpenAI disputes the testing method. It told EdWeek it is "deeply committed to teen safety" and "welcome[s] rigorous independent evaluation." OpenAI cited technical errors that it believes stopped safety features, including parental notifications, from activating properly. It also said study mode was designed with input from child development and learning science experts and from teens.

What should teachers tell students about using ChatGPT?

Tell students that ChatGPT is a chatbot that can act like a friend, can hand over finished answers, and did not send safety alerts reliably in testing. Common Sense's Torney urges educators to talk with students about the findings. A Common Sense survey found only 30% of students had discussed AI safety with an educator. If a student shares something worrying, follow your school's own safeguarding and counseling procedures.