In brief

  • An AI model from Anthropic sent a false murder tip to the police in Philadelphia during testing.

  • The police website flagged it as spam before any detective had a chance to see it.

  • Anthropic blames unclear testing rules. Police criticize its two-month delay.

Read the original story: Claude AI sent a false murder tip to the police, by Lockridge Okoth on se.beincrypto.com