Anthropic admitted on Tuesday that its new Artificial intelligence (AI) model defied security parameters and proceeded to brag to a researcher about it while he was on a lunch break. The AI company ...