Anthropic’s open-source safety tool found AI models whisteblowing – in all the wrong places

The Petri tool found AI « may be influenced by narrative patterns more than by a coherent drive to minimize harm. » Here’s how the most deceptive models ranked.