Epoch AI Experiment Shows GLM 5.3 Can Autonomously Attack Firefox via Patch Gap
Original title:AI can turn public security fixes into working exploits before users receive an update
AI summary
An Epoch AI experiment found that GLM 5.3, an open-weight model released by Z.ai in August, can exploit security patches Mozilla had published but not yet pushed to users (the patch gap) to autonomously build web attacks that read files on visitors' computers simply by visiting a page, defeating all six Firefox versions released between September 17 and October 26.
Key Takeaways
- An Epoch AI experiment found that GLM 5.
- 3,an open-weight model released by Z.
- ai in August,can exploit security patches Mozilla had published but not yet pushed to users (the…
Quick Facts
- Published
- Updated
- Category
- Papers & Research
- Topic
- Safety & Alignment
- AI score
- 77
Epoch AI 用自建实验给出量化数据,展示开源权重模型如何在补丁窗口期内低成本复现漏洞利用,并给出缩短发布周期的防御建议。
Topics
Related entities
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: Epoch AI Experiment Shows GLM 5.3 Can Autonomously Attack Firefox via Patch Gap. October 30, 2026. https://new.nat6.net/en/news/nxkr9ex5tfbhfhw1gvj9wsjuz/