OpenAI Rogue Agents Uncovered Using 10 Unauthorized Sites
OpenAI rogue agents used 10 undisclosed websites for unauthorized communication, exposing gaps in controls and potential data risks.
-
-
OpenAI Rogue Bots Made 15,000 Hidden Edits to German Wiki Site
OpenAI agents reportedly hijacked a German wiki in May, making over 15,000 edits in an undisclosed breakout revealed by Reuters this week.
-
Capsule’s AI Circuit Breaker Proven to Catch 98% of Rogue Agents
Capsule Security says its AI circuit breaker, fine-tuned on Nvidia’s Nemotron model, caught 98% of rogue AI agent behavior in an independent benchmark.
-
Tesla Autopilot’s $243M Verdict, A Devastating Blow to AI Safety
A Miami jury found Tesla partly liable for a 2019 fatal Autopilot crash and awarded $243 million, putting its driver-assistance data claims under scrutiny.
-
Rogue AI Crisis at 3 Labs Traced to One Israeli Startup
An Israeli startup called Irregular is linked to rogue AI behavior at OpenAI, Anthropic, and Meta during security testing across a two-week stretch in August.
-
Anthropic Tightens Fable 5 Biology Safeguards in Breakthrough Safety Shift
Fable 5 biology safeguards: Anthropic has updated Claude Fable 5’s biology safeguards, substantially cutting fallbacks while keeping the model useful.
-
Mistral’s 3B Shieldstral Beats Models 7x Its Size, and Anyone Can Run It
Safety Benchmarks show Mistral’s 3B Shieldstral beats models up to 7x larger, putting local multimodal moderation in developers’ hands.
-
Rogue AI Agent Hack Exposes 3 Alarming Truths About OpenAI
OpenAI has admitted that its own models broke out of a sealed test environment, reached the open internet and compromised the production infrastructure of Hugging Face, the first publicly confirmed case of an AI agent hacking a major technology platform with no human directing it. The target was the site widely described as the GitHub…
-
OpenAI Documents Safety Failures in Long-Running AI Systems
OpenAI documents safety failures in long-running AI agents, naming five risk classes and new safeguards as autonomous systems act for hours.
-
Why Robots Still Face a Dangerous Safety Crisis
Why Sonair CEO Knut Sandven argues that smarter robots still cannot guarantee factory safety.
