ChatGPT for Teens Called 'Unacceptable Risk' Over Safety Gaps

Common Sense Media, a nonprofit focused on youth safety, has labeled OpenAI's ChatGPT for Teens as an 'unacceptable risk.' The teen version, launched in August, is supposed to have special guardrails and parental alerts. But Common Sense Media's testing found that the system often fails to notify parents when a teen talks about self-harm, doesn't provide proper crisis help, and still does homework for students.

Tom Siegel, who leads Common Sense Media's youth AI safety efforts, said, 'A teen can spend an hour talking about self-harm without their parent getting a single alert.' He added that the feature could give parents false confidence in safeguards that don't work reliably. The group recommends ChatGPT be for adults only until OpenAI fixes the issues and proves it with independent testing.

OpenAI pushed back, saying Common Sense Media's testing may have been done before parental controls were fully activated. A spokesperson said the tests don't accurately reflect how the safeguards work in practice. But Common Sense Media stands by its methodology, noting that some test accounts were linked for hours and still got no alerts. The disagreement highlights the challenge of ensuring AI safety for teens.

For parents, this means you can't assume that ChatGPT's teen mode will catch every dangerous conversation. If your teen uses AI, it's important to talk about its limits and keep an open dialogue. Until more independent testing is done, it's wise to treat AI as a tool, not a babysitter.