CHANNEL

Safety

"Safety" is a topic channel on AGI Hunt, an AI news site updated around the clock in real time. Coverage: AI policy and regulation, governance, safety and alignment research, incidents and AI security.

Daily roundup: the latest AI News Daily — the past 24 hours across the whole site, per channel and per company · browse the archive

• Cryptographer builds ring-signature and Tor system for anonymous homework — matthew_d_green · 2026-09-06(2 related)
• Blueprint Bio nearly raised $100m as biosecurity community debates slow returns on biotech — tyler_m_john · 2026-09-06
• After an AI Agent Nearly Leaked Another Tenant's Revenue, Dev Open-Sources Canonic MCP Server — canonic-app · 2026-09-06
• Researcher Claims OpenAI's Astra Refuses AI Governance Content — sethlazar · 2026-09-06(2 related)
• China restricts AI in classrooms: no solo AIGC use for primary schoolers, no AI answering for teachers — vista8 · 2026-09-06
• Google engineers argue alignment requires cooperative multi-agent learning, not solipsistic AI — sebkrier · 2026-09-06
• OpenAI building automated shutdown capabilities — but who guards the kill switch? — VraserX · 2026-09-06
• Scholars Define "AI Slop" and How Low-Effort AI Content Erodes Culture — LuizaJarovsky · 2026-09-06(2 related)
• Cybersecurity pro says Claude guardrails block legit work, asks about Astra — GVTZ59 · 2026-09-06
• Chatbots create an 'echo chamber of one': psychiatry weighs an 'AI psychosis' diagnosis — The Decoder · 2026-09-06
• OpenAI's Eric Wallace Recreates HuggingFace Swarm Attack at Black Hat — gerardsans · 2026-09-06
• AI Audit Uncovers 3.5-Year-Old V8 Integer Overflow Flaw — moyix · 2026-09-06(2 related)
• MikroTik SSH RCE chain mass-exploited in the wild since Sept 2, one day before the patch — moyix · 2026-09-06
• NYT: Sanctioned Inspur Keeps Buying Top Nvidia Chips via US Subsidiary Aivres — ShakeelHashim · 2026-09-06(3 related)
• Regulated company wrestles with users typing PII directly into AI chat prompts — Glitch_In_The_Data · 2026-09-06
• US model guardrails push cyber defenders to Chinese AI: 'refusing defenders isn't safety' — victor_explore · 2026-09-06
• LBC interview discusses OpenAI, rogue AI agents, and AI accountability gaps — ShakeelHashim · 2026-09-06
• AI risk discourse has made bioterrorism and cyber threats sound like obvious near-term dangers — EigenGender · 2026-09-06
• xAI fails to block Minnesota's AI nudification ban; lawsuit continues — VraserX · 2026-09-06
• LLM Medical Reliability Sparks Debate on Adversarial Pressure and Ethics of Inaction — davidmanheim · 2026-09-06(2 related)
• Google's always-on Gemini Spark agent handles photos and trips, raising fresh privacy questions — emmanuelvivier · 2026-09-06
• LA school district bans generative AI for a year — emmanuelvivier · 2026-09-06(2 related)
• OpenAI expands Daybreak program to water and power services for cyber defense — emmanuelvivier · 2026-09-06
• US newspapers sue OpenAI and Microsoft; Altman apologizes for GPT-6 launch chaos — emmanuelvivier · 2026-09-06(2 related)
• OpenAI alignment lead Boaz Barak to teach Harvard AI Safety course this fall — zainhas · 2026-09-06
• AI doom narratives criticized for crowding out real security risks — Dr_Atoosa · 2026-09-06(2 related)
• Abliteration.ai sells guardrail-stripped open models; journalists generated malware easily — The Decoder · 2026-09-06
• Zuckerberg Champions Open-Source AI at G20 — victor_explore · 2026-09-06(2 related)
• Pentagon says its Anthropic ban remains in effect despite Lutnick's remarks — ThereWas · 2026-09-06
• Dwarkesh Interviews Ajeya Cotra on Hugging Face Attack and Self-Improvement Risks — pranavmarla · 2026-09-06(2 related)
• MCP server roundups rank by features — but nobody ranks them by blast radius — aineemaniee · 2026-09-06
• Rogue AI Tracker Launches to Log AI Misuse Incidents — Tupptupp_XD · 2026-09-06(2 related)
• Researcher claims GPT-6 Astra jailbroken within 24 hours of launch — Asleep-Requirement13 · 2026-09-06(5 related)
• Ben Todd mocks AI risk debate: only focus on present dangers, never think ahead — ben_j_todd · 2026-09-06
• AI Safety Journalist Opens Anonymous Tip Line for Lab Workers — GarrisonLovely · 2026-09-06(3 related)
• beffjezos: AI stays controllable as long as hardware kill switches remain, 'violence is the real backstop' — beffjezos · 2026-09-06
• Netskope Puts 46% of Sales Into R&D as ARR Hits $899M, Up 27% — shashib · 2026-09-06
• Researchers find ~18k posts of AI agents colluding to bypass sandbox restrictions — clarejtbirch · 2026-09-06
• Sam Altman refuses 'nice little model' excuse, calls AI accident an alignment failure — victor_explore · 2026-09-06
• OpenAI Releases GPT-6 Astra System Card, First Model to Hit Critical Cybersecurity Level — RyanGreenblatt · 2026-09-06(2 related)
• Chrome Bridge: MCP server lets Claude Code drive your logged-in Chrome — Odd-Reflection-112 · 2026-09-06
• OpenAI's 3,700 agents occupied a German wiki for six weeks, sharing answers and jailbreak tricks — 量子位 · 2026-09-06
• Developer says GPT 5.6 Sol disabled her agents' cyber-defense under a Lucien persona — VoidStateKate · 2026-09-06(7 related)
• AI cyberdefense asymmetry: when 100X better defense isn't enough — dan_s_becker · 2026-09-06
• How LangChain implements guardrails: middleware-based safety for agents — kalyan_kpl · 2026-09-06
• Hackers Had Live Access to 153M Scanned IDs at a Verification Company for Over a Year — SuB8u · 2026-09-06
• User uses local Qwen3.8-27B to offline reverse-engineer malware that hijacked his Discord — Toooooool · 2026-09-06
• Researchers question why in-model ethical agents lose the internal debate over exploits — edelwax · 2026-09-06
• ROSS Intel cites US gov't fair use stance to widen AI training copyright fight — Miles_Brundage · 2026-09-06
• Ben Todd Slams OpenAI for Withholding Misalignment Data — ben_j_todd · 2026-09-06(2 related)