英国 NCSC 发布 2026 年度评估：AI 将加速现有网络攻击手法
NCSC 年度评估说 AI 威胁是老手法变快不是新招，还点了漏洞挖掘和攻防差距，做安全的人可以看看原文。
英国国家网络安全中心（NCSC）在 2026 年度评估中给出关键判断：AI 会加速和放大现有攻击手法，不会创造全新攻击方式。报告指出 AI 模型已能在大型代码库中找到潜伏多年的漏洞，防御方需在可用 exploit 出现前抢时间修补。NCSC 还警告能使用 AI 做防御的组织与不能使用者之间差距将扩大，并点名中国最有条件利用 AI 攻击关键国家基础设施。
The @NCSC has set out its key judgements on how AI will change the cyber threat to the UK, as part of its 2026 Annual Review.
Its main call is that AI will speed up and scale existing attack methods rather than create new ones, and that human operators stay essential. Skilled attackers get the most out of it.
On vulnerabilities, it says AI models can now find flaws in large codebases that had gone undetected for years, so defenders face a race to patch before working exploits appear.
It also expects a widening gap between organisations that can use AI for defence and those that can't. And it names China as best placed to use AI against critical national infrastructure.
