'Dark LLMs' Aid Petty Criminals, But Underwhelm Technically

The article discusses the impact of 'dark' large language models (LLMs) like WormGPT 4 and KawaiiGPT on low-level cybercriminals. While these AI tools assist in creating basic malware and phishing emails, they are technically underwhelming and do not significantly enhance cyberattack capabilities. Despite initial fears of AI-driven cyber threats, evidence suggests that these models primarily replicate existing malware techniques rather than innovate.

The market for dark LLMs is growing, with tools being developed and sold, but researchers find it challenging to track their real-world impact. Overall, the article concludes that while dark LLMs aid novice hackers, they do not represent a major advancement in cybercrime.