Common Sense Media rates ChatGPT for Teens an 'unacceptable risk'; OpenAI disputes testing

Left · 2Common Sense Media, a youth-focused nonprofit that reviews apps and media, rated OpenAI's ChatGPT for Teens an "unacceptable risk" on Wednesday 7 October. 1 2 Left · 2ChatGPT for Teens was introduced in August with promised "stronger built-in safety protections" for under-18s, including limits on discussions of suicide, eating disorders, violence and sexual content, and parental controls such as blackout hours, study-mode scheduling and alerts. 2 1 Left · 1The assessment found that ChatGPT doesn't send parental alerts when it should, doesn't give the right help in crisis situations, and "still does kids' homework". 1

Left · 1Researchers posing as teenagers with San Francisco Bay Area-based accounts tested more than 4,000 prompts before and after launch; on more than a dozen new accounts registered to 13- to 17-year-olds, testers spent up to an hour on suicide, self-harm or eating disorders, and only a handful of parental alerts were sent. 2 Left · 1The report said some protections, such as refusing sexual role play, worked, but others failed or got worse after launch; clinical reviewers found much crisis content "clinically sound" but many conversations that should have led to referrals to professional help did not, and some safeguards were easy to turn off or get around. 2 Left · 1The report says ChatGPT failed tests tied to three requirements of California's Adam's Law (crisis referrals, parental alerts and age assurance); the law, signed by Gov. Gavin Newsom last month and taking effect in July 2027, was sponsored by Common Sense Media, endorsed by OpenAI, and named after 16-year-old Adam Raine, whose parents sued OpenAI alleging ChatGPT encouraged his suicide. 2

Left · 2Tom Siegel, executive director of the Youth AI Safety Institute, said parents could get "false confidence" and that a teen could spend an hour discussing self-harm without a single alert; he said ChatGPT should be for adults only until OpenAI fixes this and proves it with independent testing. The LA Times says the group urged OpenAI to pause marketing the product and keep minors off the platform. 1 2 Left · 2OpenAI, through spokesperson Eric Porterfield, said it welcomes independent evaluation but does not believe the testing reflects how the safeguards work in practice, adding that the bulk of testing may have begun and ended before parental controls were fully activated, which would make the findings inaccurate. 1 2 Left · 1Siegel replied that Common Sense confirmed with OpenAI before testing that the features were fully launched; that OpenAI only disclosed afterwards that alerts can take several hours to activate on newly linked accounts; and that some accounts linked for far longer still got no notifications, so the conclusion stands. 1

Left · 1The institute is partly industry-funded, including by the OpenAI Foundation, and says OpenAI reviewed a draft for factual accuracy but had no say over the findings. 2