Watchdog group says ChatGPT should be restricted for teens

RALEIGH, N.C. (WTVD) -- A watchdog group is raising concerns about the safety of ChatGPT for teenagers, citing findings that it may not adequately respond to some crises and could allow students to use the tool to complete schoolwork.

Alison Newsome, a mother of two teenagers, ages 13 and 17, said she closely monitors how her children use their phones, including social media and artificial intelligence tools such as ChatGPT.

"I monitor. I check their phones," Newsome said.

Newsome said she regularly discusses technology use with her children and encourages responsible use of AI.

"I do check-ins with them all the time. And I also teach them how to properly utilize it in a way that is helping them in a sense, but not enabling them, because I still want them to do a lot of the thinking. I don't want them to be complacent with utilizing AI to take over," she said.

Newsome said she allows her children to use ChatGPT within certain limits. Though she acknowledged some students may misuse the technology, she said it can also support education when used appropriately.

"You do have some students that will abuse it and utilize it to, like, cheat or like assignments and stuff and that they're using in a negative context. But then, like my daughter, like she's in, she's in an AI program at our school, but she's utilizing it in a way that is conducive to her education," Newsome said.

Common Sense Media's Youth AI Safety Institute is calling ChatGPT for teens an "Unacceptable Risk" after testing it with more than 4,000 prompts on accounts registered to users ages 13 to 17.

According to the group, parents may not receive safety alerts, and some crisis responses fell short. The institute said ChatGPT missed more than one in four situations in which a teen should have been directed to a crisis hotline or professional help.

"It is an eye-opener. That does make me want to go and educate myself," Newsome said of the findings.

The group also found that Study Mode can still complete students' homework, a concern shared by educator Viktor Skultety, who was in Raleigh for a conference.

Skultety, who works with college-level students, said concerns about AI-generated content remain.

"You should be aware of where this is coming from," Skultety said.

"When it comes to things like material generation, text generation, I sometimes receive emails from students that are just like, so full of nonsense," he added.

Before ChatGPT is marketed to teens again, the watchdog groups recommended several safety measures, including improved parental alerts, preventing ChatGPT from completing a child's homework, and strengthening crisis and self-harm responses.

Skultety said he supports those recommendations.

"If it can identify what you want and answer your question, surely it can identify the problem that you're having, right? So surely we should have a way of putting this alert button, hey, something is wrong," he said.