Anthropic defends AI despite hallucinations

Although Claude once misled users, Anthropic insists AI errors are no greater than human ones and won’t block AGI.