Bridgewater argues that AI enabled cyberattacks are spreading like a virus, with open source models helping attackers find and exploit software bugs faster than defenders can patch them. The authors cover defensive programs like Anthropic’s Glasswing, the shrinking window between a fix and its exploitation and why policy responses so far fall short.
Current Efforts Aren’t Enough to Slow the AI Cyber Virus
Bridgewater
Hudson Attar, Arham Habib
Research
7 Pages
Key Takeaways
Zero Days Doubling: Zero day exploits have roughly doubled since the first truly cyber capable open source models were released this summer.
Shrinking Exploitation Window: The average time between a vulnerability's disclosure and its exploitation has fallen from about a month to 11 days.
Elevated Attack Odds: Bridgewater is managing money based on a one in three chance of an economically destabilizing cyberattack within the next year.