On July 27, 2026 in Beijing, the China Cyberspace Security Association’s
Critical Information Infrastructure Protection Committee warned that frontier
large models such as MYTHOS have sharply lowered the bar for vulnerability
discovery and exploit validation, accelerating automation, scale and
intelligence in cyber offense and defense. The committee said faster zero‑day
discovery is making traditional notify‑and‑patch defenses inadequate and called
for stronger unknown‑threat detection, attack‑surface management,
defense‑in‑depth, behavioral monitoring and rapid response, plus broader use of
AI to improve vulnerability governance. It urged continued tracking of AI risks, tighter
integration of technology with frontline defenses and accelerated construction
of a new critical‑infrastructure security framework adapted to intelligentized
attack‑defense dynamics.
