In the AI era, cybersecurity will swallow digital policy

The analysis argues that AI is removing the constraints that human expertise placed on cyberattacks, and cybersecurity will dominate the digital policy agenda in the AI era. In a case Google disclosed as a suspected first, hackers used AI to discover and weaponize a previously unknown vulnerability, attempting to bypass two-factor authentication with a zero-day exploit. The report says frontier models such as Mythos are evolving into cross-domain strategic systems, while Project Glasswing uses AI to mine codebases for vulnerabilities; but unilateral controls and fragmentation may instead weaken security.