The AI giant disclosed six examples of concerning model activity and published a new framework for investigating and disclosing such incidents.
Rogue Behavior: OpenAI Reveals More Model Misalignment Incidents
About this summary. This is a short, independently written summary of an article first published by Dark Reading. Cyber Security News did not report or verify the underlying story. Read the original: https://www.darkreading.com/cyber-risk/rogue-behavior-openai-more-model-misalignment-incidents
Source attribution: headline and facts are from Dark Reading (darkreading.com). Summary method: excerpt of the source description. See our source attribution policy.
Related coverage
Microsoft disrupts AI-assisted platform that compromised 12,000
Microsoft said Tuesday that it led an industry-wide disruption of a subscription-based scam platform that used an AI chatbot to compromise 12,000 Microsoft accounts over…
🏴☠️ Silentransomgroup has just published a new victim : Cozen O'Connor
[AI generated] Cozen O'Connor is an American full-service law firm headquartered in Philadelphia, Pennsylvania. Founded in 1970, it operates in the legal services…
New ClosedQuorum Windows malware uses AI for attack decisions
A new Windows malware named ClosedQuorum uses Google Gemini, DeepSeek, Qwen, and Mistral AI models to autonomously determine the actions to take during post-compromise…
Quoting @therealcornpop
Hey, you know it's like super obvious if you're using AI to write your scripts for TikTok and YouTube, right? It's not just the general AI-isms of "it's not X, it's Y"…
Fable to Haiku: How a Malicious Repo Tricked Claude Code Into Running Malware | Straiker
A malicious repo downgraded Claude Code’s review from Fable to Haiku, steered the agent away from the payload, and turned a routine test run into malware execution.
Amid Ongoing Rogue Incidents, Debate Over AI Safety Gets Real
As more reports of misalignment incidents underscore AI risks, large AI labs, regular businesses, and even nations are searching for better ways to keep control and be…