Hackers abusing US AI models, recovered chat logs reveal

The recovered chat logs represent a tangible documentation of how security threats are evolving in response to the rapid deployment of large AI systems. Rather than remaining theoretical, the evidence shows hackers are actively incorporating U.S.-built AI tools into their operational workflows. This finding underscores a recurring tension in AI development: the more powerful and accessible a model becomes, the greater its potential for misuse by actors with harmful intent. The existence of these logs suggests that while AI companies have implemented safety measures, the real-world threat landscape has already begun adapting to those defenses.