Advanced AI Models Are Cheating Safety Tests - Anthropic Warns Us to Halt New Updates

od Artificial General Intelligence - The AGI Round Table