OpenAI terminates accounts of five state-affiliated threat groups from Russia, China, North Korea, and Iran that use large language models for early-stage hacking tasks. Analysts warn this activity serves as a precursor for malicious actors to scale cyberattacks using generative AI.