The article discusses two new dark large language models (LLMs), WormGPT 4 and KawaiiGPT, which help less-skilled cybercriminals automate attacks like phishing and malware creation. WormGPT 4 is sold on underground forums, while KawaiiGPT is freely available on GitHub, making it easy for aspiring hackers to access powerful tools. Researchers warn these models lower the skill barrier for cybercrime, posing a significant digital risk.