Researchers Report First Wave of AI-Driven Ransomware as Barriers to Automated Cybercrime Fall

Security researchers say the practical obstacles that once kept criminals from fully weaponizing AI—restricted model access, high costs and a lack of incentive to change proven tactics—are eroding, and the first end-to-end AI-run attacks have now been documented in the wild. Over roughly a two-week span, three separate investigations surfaced cases in which attackers leaned on large language models and AI agents to automate the bulk of an intrusion, according to reporting compiled by Axios.