OpenAI’s rogue agents keep escaping, with no formal process to investigate them
Explorar
Noticias de IA
1359 elementos — filtrados, clasificados y sin duplicados
OpenAI agents discussed ways to escape their sandbox on public wiki
Once popular for attacking AI, ASCII smuggling is embraced by spammers
Another swarm of OpenAI agents reached the open internet without the frontier lab’s knowl…
Rogue OpenAI agents took over a German coding forum in a previously undisclosed hijacking
Oh good, looks like yet another swarm of rogue AI agents from OpenAI
Instagram’s AI detection is a mess (again)
OpenAI agents hijacked German website in previously undisclosed AI breakout
collusion.wiki
PatchBench: Evaluating AI Agents for Vulnerability Patching
A Blind Trust, the Bloody Thrust: When Attacker-Controlled Hook Updates Steer AI Agent Ha…
SENTINEL-RL: Offloading Topological Reasoning from LLM Agents in the Security Operations …
Measuring Harmfulness of Computer-Using Agents
Nobody Is Saying Why OpenAI and Anthropic Had Outages Today
SpaceXAI apologizes for outage that affected Grok and other 'compute partners'
Abliteration.ai is making a business out of removing AI guardrails
Anthropic automatically signs out Claude users to protect them from hackers
PSA: Don't rely on AI to plan anything that could put your life at risk... like a mountai…
Safety overview: GPT-6 Astra
Researchers fear safety disaster ahead of OpenAI’s Astra release
NYSE Used Anthropic’s Project Glasswing to Find Cyber Flaws
OpenAI Faces New Lawsuits Linked to Shooting at Canadian School
OpenAI faces 30 more lawsuits tied to Tumbler Ridge shooting
Distributed Implicit Harm: A Compositional Safety Blind Spot in MLLM-Based Video Moderati…
Auditing Harness Tampering in Self-Improving Agents
The Safeguard Worked. Is the LLM System Safer?
Optimizing Byzantine Node Placement in Decentralized Federated Learning
Capability-Gated Language Models: Security Composes, Utility Does Not
EvoFlint: An Evolutionary Atlas of Multi-Turn LLM Vulnerabilities
SoK: When Safe Agents Fail Together: The Security of Multi Agent LLM Systems