English(EN) Using AI to patch a vuln. generates Fix-Like Artifacts with Embedded Defects (FLAWED). In security research, our job is to ask hard questions and then find out

研究发现AI生成的代码补丁成功率仅为26%

作者 PulseAugur 编辑部 · [1 个来源] · 2026-08-06 21:35

来自1Password的Off-by-1 Labs的研究人员发现，AI生成的代码补丁成功率仅为26%。此外，这些成功补丁中有三分之一是脆弱的，可能无法经受进一步审查。该团队开源了一个工具并发布了与其发现相关的数据集，这些发现已在Black Hat和DEF CON上发表。 AI

影响 AI生成的代码补丁被证明是不可靠的，这表明在软件安全方面仍需要人工监督。

排序理由该集群报道了一篇已发表的研究论文和一个关于AI生成代码补丁有效性的开源工具。[lever_c_demoted from research: ic=1 ai=1.0]

在 Mastodon — sigmoid.social 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

报道来源 [1]

Mastodon — sigmoid.social TIER_1 English(EN) · [email protected] · 2026-08-06 21:35

Using AI to patch a vuln. generates Fix-Like Artifacts with Embedded Defects (FLAWED). In security research, our job is to ask hard questions and then find out

Using AI to patch a vuln. generates Fix-Like Artifacts with Embedded Defects (FLAWED). In security research, our job is to ask hard questions and then find out the answers. Today, it’s time to Find Out. I’m thrilled to share our first piece of research from Off-by-1 Labs at 1Pass…

链接 1password.com/…/why-ai-generated-patches-…

报道来源 [1]

Using AI to patch a vuln. generates Fix-Like Artifacts with Embedded Defects (FLAWED). In security research, our job is to ask hard questions and then find out

相关实体

相关话题