CyberScoop reports that a sophisticated underground market for AI-powered hacking tools is rapidly emerging, lowering the barrier to entry for cybercrime, according to a new report from Palo Alto Networks' Unit 42.
Researchers have identified the sale of custom, jailbroken, and open-source large language models on dark web forums, which are specifically designed to assist with tasks like vulnerability scanning, data encryption, and writing malicious code. Andy Piazza, a senior director at Unit 42, explained that this mirrors the history of legitimate security tools being co-opted by criminals, stating, And now were seeing the same thing with AI.
The report highlights two key examples: a new commercialized version of WormGPT, available via cheap subscription, and a free, user-friendly model called KawaiiGPT. While the code these models generate is often easily detectable, their true danger lies in democratizing cybercrime.
Piazza emphasized that these tools provide "interoperability," allowing unskilled individuals to generate attack scripts using simple, conversational language, thereby significantly lowering the technical expertise required to launch cyber attacks.
The full story
This article is one source in a clustered incident — the cluster page carries the summary, timeline and every other outlet covering it.
