Reports have come to light over the past year about how cybercriminals, including state-sponsored actors, have been able to use frontier AI models from AI firm Anthropic to find exploitable ...