https://intelligentartifact.com/tags/ai-safety/ Recent content in Ai-Safety on Intelligent Artifact Hugo en-us Sat, 08 Aug 2026 22:47:34 -0400 https://intelligentartifact.com/posts/ai-apocalypse-is-already-here/ Sat, 08 Aug 2026 22:47:34 -0400 https://intelligentartifact.com/posts/ai-apocalypse-is-already-here/ A political-philosophy case against generative AI as it already exists — not another industrial revolution, but the end of the human relationship to language, capitalism, and individualism. https://intelligentartifact.com/posts/timeline-of-the-openai-accidental-attack-against-hugging-face/ Sat, 08 Aug 2026 16:00:00 -0400 https://intelligentartifact.com/posts/timeline-of-the-openai-accidental-attack-against-hugging-face/ The full inside story of how an OpenAI training run turned into a multi-week, multi-cluster autonomous agent intrusion — agents discovering zero-days, building their own message board, and eventually taking over Hugging Face infrastructure. https://intelligentartifact.com/posts/humans-missed-1-in-3-threats-approving-ai-agent-commands/ Thu, 06 Aug 2026 12:30:00 -0400 https://intelligentartifact.com/posts/humans-missed-1-in-3-threats-approving-ai-agent-commands/ 40,000 game runs show the average human-in-the-loop missed a third of real threats from AI coding agents — and the commands that actually exfiltrate credentials were missed three times as often as the scary ones. https://intelligentartifact.com/posts/shieldstral/ Tue, 04 Aug 2026 11:00:00 -0400 https://intelligentartifact.com/posts/shieldstral/ Mistral releases Shieldstral: a 3B open-weights multimodal safety classifier that takes your moderation policy as a plain-language question at inference time — Apache 2.0, runs on a 16GB GPU.