John Hultquist, a principal analyst at Google's Threat Analysis Group, stated that hacker attacks targeting AI accounts and computing resources have significantly increased this year, with a new attack method called "LLM hijacking" gaining traction in the cybercrime underground. Access to AI models from companies like Anthropic, Google, and OpenAI is being sold on dark web markets at discounts of up to 97%. Simultaneously, criminal gangs and state-sponsored hacker groups are infiltrating enterprise cloud servers to implant their own AI models, "freeloading" hashrate in a manner identical to previous cryptocurrency mining hijackings. Hultquist warned that enterprises are more vulnerable to attacks during the initial deployment of AI infrastructure, and defenders will face a "cost asymmetry" dilemma.