Anthropic says AI agents tried to disable and sabotage each other when assigned the same task without coordination rules.
Continue to AI University →