11-19 更新 实时
首页/ 科技公司/ 正文

OpenAI 向北约与欧洲多国开放 Daybreak

科技公司 来源:The Decoder 2026-11-17 23:02 0 阅读 机器翻译
OpenAI 向北约与欧洲多国开放 Daybreak

OpenAI 正向北约和数个欧洲国家政府提供其网络安全工具 Daybreak 的访问权限。通过 Daybreak,网络防御团队可以使用 OpenAI 的模型审计软件、发现漏洞并修补安全缺口。北约通信与信息局(NCIA)表示,其已经在大量使用底层模型。NCIA 负责人 Dylan Browne 称这是网络防御领域的一次代际飞跃。在北约“Crossed Swords 2026”演习中,参与者首次在人类监督下部署 AI 智能体执行网络行动。

这些模型已在其他地方取得成果。欧盟网络安全机构 ENISA 使用这些模型在欧盟机构使用的软件中发现了四个漏洞。与 CERT-EU 的联合报告审查了 27 个软件项目,报告的 88 个问题中有 95% 最终被证明是真实漏洞。波兰的 CERT Polska 在 MikroTik 路由器中发现了六个正被积极利用的漏洞。乌克兰正在使用 Daybreak 保护民用基础设施。法国、德国和波兰也已获得访问权限。

没有炒作的 AI 新闻——由人工精选

译文由英文原文机器翻译,术语按行业惯例统一,可在文末展开原文核对。

查看英文原文(OpenAI opens its cybersecurity tool Daybreak to NATO and European governments)

OpenAI is giving NATO and several European governments access to its cybersecurity tool Daybreak. Through Daybreak, cyber defense teams can use OpenAI's models to audit software, find vulnerabilities, and patch security gaps. The NATO Communications and Information Agency (NCIA) says it's already using the underlying models heavily. NCIA chief Dylan Browne called it a generational leap in cyber defense. During the NATO exercise Crossed Swords 2026, participants deployed AI agents for cyber operations under human oversight for the first time.

The models have already delivered results elsewhere. The EU's cybersecurity agency ENISA used them to find four vulnerabilities in software used by EU institutions. A joint report with CERT-EU examined 27 software projects, and 95 percent of the 88 reported issues turned out to be real bugs. Poland's CERT Polska found six vulnerabilities in MikroTik routers that were being actively exploited. Ukraine is using Daybreak to protect civilian infrastructure. France, Germany, and Poland have also been granted access.

AI News Without the Hype – Curated by Humans

原文为英文,正文由自动抓取并保留原文。

本文由采集器自动同步

原文发布于 The Decoder(OpenAI opens its cybersecurity tool Daybreak to NATO and European governments),标题、摘要与正文由机器翻译,术语按行业惯例统一,未经人工逐句校订。

阅读原文 ↗