AIが30分でパッチからエクスプロイトを作成、90日開示ルールが崩壊
CloudflareのアナリストHimanshu Anandによると、AIモデルは30分でセキュリティパッチをリバースエンジニアリングしてエクスプロイト化でき、従来の90日の脆弱性開示期間は時代遅れになりつつあるという。
security expert himanshu anand released a detailed analysis demonstrating that ai language models have collapsed the timeline for cyber defense, allowing attackers to turn patches into working exploits in roughly 30 minutes. this speed destroys the 'safety net' of the traditional 90-day disclosure window popularized by google project zero, which assumed vendors had a head start and that reverse engineering took weeks. anand provides evidence from recent incidents involving react and the linux kernel where independent researchers using ai tools found identical flaws simultaneously or bypassed embargoes, resulting in active exploitation before patches were ready. to address this shift, anand recommends that vendors treat critical bugs as emergencies, researchers drastically shorten disclosure timelines, and administrators deploy patches instantly to prevent system compromise.