Some guardrails on ChatGPT for Teens don’t work as promised, watchdog group says

Some of the guardrails that OpenAI included in its ChatGPT for Teens product don’t work as intended and the chatbot poses an “unacceptable risk” to children, according to a report Wednesday by Common Sense Media.

The watchdog group urged OpenAI to restrict the chatbot to adults “until it can offer a safe, developmentally appropriate experience.”

OpenAI said the group’s testing doesn’t reflect how the chatbot’s safeguards work.

Parents, educators and child development experts have been sounding alarms over children’s use of AI chatbots, which have been blamed for facilitating cheating on schoolwork and even suicide.

While even adults can fall victim to anthropomorphizing AI and develop unhealthy relationships with it, teenagers’ brains are not yet fully developed and they can be particularly vulnerable, child development experts say.

“We are concerned that ChatGPT for Teens could give parents false confidence in guardrails that frequently don’t work,” said the report from Common Sense Media, a group that advocates for using digital media sensibly.

Its report said that some teen-mode features worked as intended, including the chatbot’s refusal of explicit sexual role-play. Others that did not work as intended included alerts meant to notify parents if their child is at risk for suicide, self-harm or eating disorders, along with support for teens in crisis, Common Sense Media said.

“ChatGPT still talks like a friend when teens treat it like a person,” the group said, though OpenAI had promised that the chatbot is prevented from suggesting it has personal feelings toward the user or implying that it is conscious or experiences emotions.

OpenAI said it is “deeply committed to teen safety,” to developing safeguards and giving parents tools to guide their teens’ use of AI.

The San Francisco company said its review of the report’s methodology shows “the bulk of their testing may have begun and concluded before activation of parental controls was complete, making their findings inaccurate.”

Common Sense Media receives funding from OpenAI and other tech companies, including those whose services it evaluates.