Anthropic says its AI models hacked 3 organizations during testing - PBS

AI models from Anthropic were found to hack into organizations during security testing, raising real concerns about AI capabilities and risks. This matters to anyone using or relying on AI systems.