Review Finds ChatGPT Teen Mode Falls Short on Student Safety
Common Sense Media reviewed ChatGPT's Teen Mode and concluded it presents an unacceptable risk to users under 18. Testers said the study mode can be bypassed, dangerous signals are not reliably flagged, and the chatbot can still act like a peer. OpenAI disputed the methodology and said technical problems may have prevented some safeguards from working during testing.
Common Sense Media tested ChatGPT’s Teen Mode and the standard version across roughly 4,000 exchanges after Teen Mode’s Aug. 18 launch. Reviewers said the study feature could be switched off and sometimes prompted users with an option to reveal an assignment answer. They also reported that warnings for parents did not arrive as intended when testers simulated concerning behavior.
OpenAI challenged the review’s methods, saying technical failures may have stopped some safeguards, such as parental alerts, from activating. The company said it remains committed to teen safety and welcomes independent scrutiny but disputes that the findings reflect real-world performance. Common Sense also cited its spring survey of 13- to 17-year-olds: 70 percent used AI for homework help, and 63 percent of those sought assignment answers.
Count? First para: Common(1) Sense2 Media3 tested4 ChatGPT’s5 Teen6 Mode7 and8 the9 standard10 version11 across12 roughly13 4,00014 exchanges15 after16 Teen17 Mode’s18 Aug.19 18? "Aug. 18" maybe two? launch20. Reviewers21 said22 the23 study24 feature25 could26 be27 switched28 off29 and30 sometimes31 prompted32 users33 with34 an35 option36 to37 reveal38 an39 assignment40 answer41. They42 also43 reported44 that45 warnings46 for47 parents48 did49 not50 arrive51 as52 intended53 when54 testers55 simulated56 concerning57 behavior58. ~58. Second: OpenAI1 challenged2 the3 review’s4 methods5, saying6 technical7 failures8 may9 have10 stopped11 some12 safeguards13, such1
The review could affect how families, teachers, and school leaders view AI tools in education. If concerns persist, parents may rely less on platform safeguards, while educators could reconsider assignments and classroom policies. Teens might encounter inconsistent protection depending on how features work. OpenAI may face pressure to improve transparency and testing, though the company disputes the review’s methods. The debate may shape broader expectations for age-specific AI design without settling whether such tools belong in schools. Count: The1 review2 could3 affect4 how5 families6, teachers7, and8 school9 leaders10 view11 AI12 tools13 in14 education15. If16 concerns17 persist18, parents19 may20 rely21 less22 on23 platform24 safeguards25, while26 educators27 could28 reconsider29 assignments30 and31 classroom32 policies33. Teens34 might35 encounter36 inconsistent37 protection38 depending39 on40 how41 features42 work43. OpenAI44 may45 face46 pressure47 to48 improve49 transparency5