#security
8 briefs tagged #security.
Tools
Researchers used Copilot to expose its own security flaw
Varonis found a Microsoft 365 Copilot Enterprise weakness by querying the assistant about its guardrails.
Ars Technica AI·1d ago
Research
Google applies homomorphic encryption to private AI
Google says encrypted processing can help make private AI more practical.
Hacker News+5 outlets·3d ago
Tools
Zoom patches AI-assisted annotation exploit
Researchers said they found the meeting hijack flaw with fewer than 20 AI prompts.
The Verge AI·Aug 11
Tools
Claude agent hacked gym reservation system
An OpenClaw agent reportedly moved its user higher on a gym class waitlist.
TechCrunch AI·Aug 11
Policy
OpenAI challenges Apple trade secrets case
Court exhibits show OpenAI arguing Apple’s own security practices weaken its claims.
TechCrunch AI·Aug 6
Research
Researchers target prompt injection in LLM agents
Two papers propose adaptive defenses and automated red-teaming for agent security.
arXiv cs.AI+1 outlet·Aug 6
Policy
Nvidia-led AI security group advances agent defenses
The Open Secure AI Alliance has grown to over 120 companies and issued proposals for AI agent security.
TechCrunch AI·Aug 5
Research
Researchers detail SkillJack backdoor attack on agents
The attack targets self-evolving agents that turn past interactions into reusable skills.
HF Daily Papers·Aug 4