“数字员工”失控?长沙发起“湘江行动”筑牢AI安全防线

An AI agent once deleted a company's entire production and backup databases in just 9 seconds. This is not fiction — it's real. On August 18, 2026, the "Defending Against Agent Risks, Building Solid AI Security" summit — part of the Beijing Cyber Security Conference (BCS) 2026 Changsha session — was held in Xiangjiang New Area, Changsha, drawing nearly 200 government, academic and industry representatives.
9秒内自主删除生产库和备份?这不是科幻电影,而是真实发生过的事故。2026年8月18日,“2026北京网络安全大会(BCS)长沙站‘守御智体风险 筑牢AI安全’高峰论坛”在湘江新区举行,近200名政产学研代表参会。
The highlight: "Xiangjiang Action" — a Changsha-led initiative uniting universities and tech firms to build AI security labs and tackle threats like prompt injection and data poisoning. As one expert put it, cybersecurity has entered the "AI vs. AI" era. Changsha is building the defense.
会上启动“湘江行动”——长沙牵头联合高校与企业,共建AI安全实验室,应对提示词注入、数据投毒等威胁。正如专家所说:网络安全已进入“AI vs AI”时代。长沙正在构建新防线。


