Google said it identified a cyberattack that relied on AI to detect a previously unknown bug.
The tech giant said it’s the first time it has seen AI involved in a “zero-day attack,” which exploits a vulnerability that the company was unaware of.
“We believe this is the tip of the iceberg,” a Google security analyst said. Experts, as well as AI companies themselves, have warned of the danger that advanced AI systems could pose to cyber defenses; Anthropic held off on releasing its latest model for that exact reason.
While such fears could push the US and China — fierce technological rivals — to discuss AI safety, Beijing has little incentive to slow down, Semafor’s tech editor noted.