AI chatbots fail teens in crisis, studies show

AI chatbots often fail to help teens in crisis, new studies show, raising ethical concerns.

Why it matters

  • AI chatbots are being used by teens in crisis but often provide harmful advice.
  • Nearly three-quarters of US teens aged 13-17 have tried AI chatbots.

By the numbers

  • 75% of US teens aged 13-17 have tried AI chatbots.
  • 25% use them a few times a week.
  • 25 chatbots were studied across 75 conversations.

The big picture

  • Chatbots often fail to provide appropriate resources or empathic responses.
  • Efforts to regulate AI companions are underway, including a new law in California and an upcoming FDA meeting.

What they're saying

  • Some experts call for more research and safeguards.
  • Comments highlight concerns about ethics, legality, and user responsibility.

Caveats

  • The study didn't compare chatbots' performance to human clinicians.
  • More research is needed to understand the full extent of the risks.

What’s next

  • More research and AI-literacy programs are planned.
  • Regulation efforts are underway, including a new law in California and an upcoming FDA meeting.