Artificial Intelligence
Common Sense Media calls ChatGPT for Teens unsafe
Common Sense Media said OpenAI’s teen version of ChatGPT failed to protect young users properly, while OpenAI argued the group’s testing did not reflect how its safeguards worked.
- Common Sense Media said ChatGPT for Teens was an unacceptable risk.
- Its assessment said the product did not issue parental alerts when needed, handled crisis situations poorly and still completed schoolwork.
- OpenAI said the evaluation may have taken place before parental controls were fully activated on newly linked accounts.
- Tom Siegel said the group stood by its testing and said some linked accounts still received no alerts even after longer periods.
Common Sense Media said OpenAI's teen version of ChatGPT failed its safety expectations. The nonprofit, which reviews apps and media for young users, said the feature did not meet what OpenAI had promised.
The assessment said the system missed parental alerts, did not cope well in crisis situations and still helped with homework. Tom Siegel said a teenager could discuss self-harm for an hour without a parent being notified.
OpenAI disputed the findings through spokesperson Eric Porterfield. He said the company welcomed independent review, but argued that Common Sense Media's tests may have begun before parental controls were fully active, which would mean they could not show whether notifications worked as intended.
Siegel said Common Sense Media had checked with OpenAI before testing and had been told that teen-mode features were live. He said some test accounts remained silent on alerts even after being linked for longer than the activation window OpenAI later described.
Named in this story
People
- Tom Siegel
- criticised the product for Common Sense Media
- Eric Porterfield
- gave OpenAI's response to the assessment
Companies
- OpenAI
- introduced ChatGPT for Teens and disputed the findings
Organisations
- Common Sense Media
- published the assessment of ChatGPT for Teens
Products and systems
- ChatGPT for Teens
- the teen-focused version of ChatGPT under review
How the source tells it
The piece reads as a safety dispute, with strong risk language from Common Sense Media and a point-by-point rebuttal from OpenAI.
- alarm a severe risk label and a self-harm example in the assessment
- conflict the story is structured around duelling statements from the two sides