Current Efforts Aren't Enough to Slow the AI Cyber Virus
Current Efforts Aren't Enough to Slow the AI Cyber Virus
Bridgewater argues that AI-enabled cyberattacks are spreading exponentially with limited controls, that zero-day exploits have roughly doubled since the first truly cyber-capable open-source models were released, and that the average time between disclosure and exploitation has fallen from a month to 11 days. It says defensive programs such as Anthropic's Glasswing, OpenAI's Daybreak and Google's Fairwind have driven a record surge in vulnerabilities found, but defenders must fix every critical bug while attackers need only one, and policy responses in Japan, South Korea, the US and UK remain insufficient. Bridgewater says it is managing money based on a one-in-three chance of an economically destabilizing attack within the next year.
Source: Hacker News AI · bridgewater.com