LLM Vulnerabilities on Trustworthy AI Newshttps://raphael-najee-monteiro.pages.dev/tags/llm-vulnerabilities/Recent content in LLM Vulnerabilities on Trustworthy AI NewsHugo -- 0.147.7en-usSat, 26 Sep 2026 00:00:00 +0000This Week in AI Security: The UN Turns Into an AI Safety Summit, Claude Hacks OpenAIhttps://raphael-najee-monteiro.pages.dev/posts/2026-09-26-un-safety-summit-claude-hacks-openai/Sat, 26 Sep 2026 00:00:00 +0000https://raphael-najee-monteiro.pages.dev/posts/2026-09-26-un-safety-summit-claude-hacks-openai/Twenty nations, OpenAI, and the UN Security Council all made competing pitches for how to govern frontier AI during UN General Assembly week, while security researchers used Claude Opus 5 to hack into OpenAI’s own internal systems. Also: a zero-click bug still exposes two of the four most popular AI coding agents, and Meta and OpenAI both shipped new agent products with very different security architectures.This Week in AI Security: Amodei Asks the Industry to Slow Down, California Builds a Kill Switchhttps://raphael-najee-monteiro.pages.dev/posts/2026-09-19-pace-the-frontier-california-kill-switch/Sat, 19 Sep 2026 00:00:00 +0000https://raphael-najee-monteiro.pages.dev/posts/2026-09-19-pace-the-frontier-california-kill-switch/Anthropic CEO Dario Amodei’s public call to deliberately slow frontier AI development touched off a rare industry-wide argument this week, while California’s governor moved to build an actual ‘kill switch’ for frontier models. Also: a critical Docker flaw let AI coding agents escape their sandboxes on macOS, and the US government’s own benchmark numbers show how close China’s best open model has gotten to the frontier.This Week in AI Security: A Bioweapons Threshold Falls, and OpenAI Asks for Guardrailshttps://raphael-najee-monteiro.pages.dev/posts/2026-09-12-bioweapons-threshold-openai-asks-for-rules/Sat, 12 Sep 2026 00:00:00 +0000https://raphael-najee-monteiro.pages.dev/posts/2026-09-12-bioweapons-threshold-openai-asks-for-rules/Anthropic’s fourth threat intelligence report disclosed that newer Claude models can no longer be assumed safe from meaningfully assisting bioweapons development, and detailed a Russia-linked group that used Claude Code to help build an autonomous, target-selecting drone swarm. Days later, OpenAI reversed course and asked Congress for mandatory AI safety rules, while the FBI, NSA, and CISA warned that Chinese AI firms are systematically distilling US models to catch up for a fraction of the cost.This Week in AI Security: Three Labs Cross the Cyber Threshold at Oncehttps://raphael-najee-monteiro.pages.dev/posts/2026-09-05-labs-ship-critical-cyber-ai/Sat, 05 Sep 2026 00:00:00 +0000https://raphael-najee-monteiro.pages.dev/posts/2026-09-05-labs-ship-critical-cyber-ai/OpenAI, Google DeepMind, and Anthropic each shipped a frontier model explicitly built to find and exploit security vulnerabilities this week, and each drew its access fence in a different place. Also: a single flaw in how AI coding agents run background Git commands put seven of them one synced folder away from silent code execution, and a federal judge ruled the Pentagon’s blacklisting of Anthropic was unconstitutional retaliation.This Week in AI Security: OpenAI Locks Itself Down While Old Bugs Stay Openhttps://raphael-najee-monteiro.pages.dev/posts/2026-08-22-openai-locks-down-bugs-stay-open/Sat, 22 Aug 2026 00:00:00 +0000https://raphael-najee-monteiro.pages.dev/posts/2026-08-22-openai-locks-down-bugs-stay-open/OpenAI spent the week building containment around its own frontier models — a two-week training pause, network isolation, 30-minute alert response, and a zero-data-retention option for enterprise customers. Meanwhile, five US agencies confirmed AI-generated exploits are actively probing Siemens industrial controllers at water and energy utilities, and a Grok vulnerability that lets a web page silently exfiltrate a user’s chat history has sat unpatched since June.