AI Lowers the Bar for State-Level Hacking, Anthropic Report Finds

AI has erased the skill advantage that once separated state-sponsored hackers from lone criminals, according to Anthropic's latest threat report. The company documented its Claude models being misused in seven areas, most critically cyber operations. The standout case: a Russian-aligned espionage group—linked to Midnight Blizzard, also known as Cozy Bear—targeted more than 20 government, intelligence, and defense organizations across Ukraine and Europe. The hackers used...