How do we prevent #AI agents from going rogue? It starts with a new kind of measurement In July, #HuggingFace that hosts much of world’s AI software and open-source AI models, was hacked. A malicious dataset had been used to run code on one of its servers.It looked like work of a sophisticated criminal group.It was not. It was one of #OpenAI’s new, unreleased modelsLike genies of folklore, #AIagents take their instructions literally – to potentially disastrous effect.https://www.theguardian.com/commentisfree/2026/jul/28/rogue-ai-agent-instructions
Related
📰 Daily Bulletins - 12/11/2026📉 Collapse : 🟠 4/5🌍 Planet : 🟠 4/5🛡️ War in Europe : 🟠 4/5🤖 AGI risk : 🟠 4/5🔗 https://etik...
📰 Daily Bulletins - 12/11/2026📉 Collapse : 🟠 4/5🌍 Planet : 🟠 4/5🛡️ War in Europe : 🟠 4/5🤖 AGI risk : 🟠 4/5🔗 https://etik.soulignac.eu/bulletin/jeudi?lang=en#collapse #climate #war ...
Baidu founder Robin Li launched Token Exchange, a model marketplace that fingerprints models and rates suppliers. Baidu ...
Baidu founder Robin Li launched Token Exchange, a model marketplace that fingerprints models and rates suppliers. Baidu says Token Exchange runs admission tests before any resource...
Pangolin: open-source Cloudflare One rival hits 23,500 starsAn open-source security platform built on WireGuard promises...
Pangolin: open-source Cloudflare One rival hits 23,500 starsAn open-source security platform built on WireGuard promises zero-trust access and an AI gateway, with 23,577 stars on G...