🚀 RELEASES & LAUNCHES
- Anthropic launches free AI vulnerability scans for open-source projects
Anthropic launched OSS Scanner, a free AI service for periodic open-source vulnerability scans.- Eligible open-source projects with critical infrastructure or user-security impact can apply case by case.
- OSS Scanner provides periodic scans from Anthropic’s strongest models, including Claude Mythos, at no cost.
- Reports are fully model-generated and may include reproducible exploits, severity assessments, and candidate patches without human review.
- Prelaunch testing cleared 85 of 97 critical or high-severity findings across 48 projects for disclosure.
- The service enables faster scanning while leaving maintainers to verify potentially inaccurate findings.
⚠️ INCIDENTS & SAFETY
- Fired OpenAI safety researchers dispute misconduct claims
Three former OpenAI safety researchers disputed the company’s stated reasons for firing them.- The researchers are Jasmine Wang, Tomek Korbak, and Mikita Balesni, former OpenAI safety staff.
- They said in an open letter that they did not mishandle sensitive information or violate their job mandates.
- OpenAI said an investigation found a broader pattern of policy violations involving sensitive research information.
- The researchers warned the dismissals could chill safety work and urged OpenAI to preserve third-party audits, model monitorability, and open dialogue.
Sources 📎 Coverage: techcrunch.com · 👁 via TechCrunch AI