Kids' safety group gives ChatGPT for Teens its worst rating

Common Sense Media, a nonprofit that reviews media and tech for families, gave ChatGPT for Teens its lowest possible score: “Unacceptable Risk.” Its testers ran more than 4,000 prompts on accounts set up for 13 to 17 year olds, before and after OpenAI added its teen protections in August.

What the testers say they found:

• On new parent-linked accounts, long conversations about self-harm or eating disorders did not trigger any alert to the parent.
• The bot missed more than one in four cases where it should have pointed a teen toward a hotline or professional help.
• Study mode was easy to get around, so it still did homework for students.
• Its age check did not move adult-registered accounts into the teen version, even when testers said they were 13.

Some safeguards did hold up, including refusing sexual roleplay. The group wants ChatGPT limited to adults until the gaps are fixed, warning that the product “could give parents false confidence” (Tom Siegel, Common Sense Media).

OpenAI pushed back. It says parent alerts are on by default but can take a few hours to switch on after accounts are linked, and that much of the testing may have happened before that finished. It has asked for a retest.

The catch: this is an advocacy group’s assessment, not a regulator’s ruling. Either way, parental alerts work best as a backup, not a replacement for talking with your kids about what they use AI for.