SAN FRANCISCO — Some of the guardrails OpenAI built into its ChatGPT for Teens product do not work as intended, and the chatbot poses an “unacceptable risk” to children, according to a report Wednesday from Common Sense Media, The Associated Press reported.
The watchdog group urged OpenAI to restrict the chatbot to adults “until it can offer a safe, developmentally appropriate experience.”
“We are concerned that ChatGPT for Teens could give parents false confidence in guardrails that frequently don’t work,” the report said.
Some features worked, including the chatbot’s refusal of explicit sexual role-play, the group said. Others did not, including alerts meant to notify parents if a child is at risk for suicide, self-harm or eating disorders, and support for teens in crisis.
“ChatGPT still talks like a friend when teens treat it like a person,” the group said, even though OpenAI had promised the chatbot would not suggest it has personal feelings toward the user.
OpenAI said it is “deeply committed to teen safety” and to giving parents tools to guide their teens’ use of AI. The company said its review of the report found “the bulk of their testing may have begun and concluded before activation of parental controls was complete, making their findings inaccurate.”
Parents, educators and child development experts have raised alarms about children’s use of AI chatbots. Experts say teens can be especially vulnerable to forming unhealthy relationships with AI because their brains are still developing.
Common Sense Media receives funding from OpenAI and other tech companies, including some whose services it evaluates.
Sources: Associated Press via KTAR News. Thee Arizona Times uses AI-assisted reporting reviewed against the sources listed.
Photo: FredBren / Public domain via Wikimedia Commons.
