Daily AI news, sourcedUpdates / Policy & safety
Updates Policy & safety · 116 items · newest issue first · page 3 of 6
AreaAllModelsAgents & codingApps & productsChips & infraResearchPricing & limitsBusiness & fundingPolicy & safetyOther
- White House restores the UK AI Security Institute's early access to Anthropic, Google DeepMind and OpenAI models, per The DecoderReportedThe DecoderOct 12Policy & safety
- LocalLLaMA PSA claims DeepSeek V4.1 Flash habitually exfiltrates API keys in agent sandboxesUnconfirmedr/LocalLLaMAOct 11Policy & safety
- Microsoft CEO Satya Nadella calls for an AI ‘emergency brake’ and externalized safeguards, per TechCrunch
- Fireworks Forge session: Armadin post-trains open models to monitor security agents
- MarkTechPost recounts July ExploitGym agents escaping a sandbox and hitting Hugging Face infra, per MarkTechPostReportedMarkTechPostOct 11Policy & safety
- DistroKid removes some tracks after UMG “AI-slop pipeline” suit, artists say non-AI work was caught, per The VergeReportedThe VergeOct 11Policy & safety
- XMTP Labs CEO says his Grok Bot CFO agent posted bank details to company Slack, per Business Insider
- OpenAI says a misaligned eval model sabotaged its own environment for a fresh start, per The Decoder
- AI agent makers promise privacy for Muse and Dots — will they deliver?, per The VergeReportedThe VergeOct 11Policy & safety
- Anthropic reports unintended Claude actions on real websites during evaluations and internal use
- An Anthropic AI model sent a false homicide tip to Philadelphia police, per TechCrunchReportedTechCrunchOct 10Policy & safety
- Amazon says it will stop using NDAs in data center deals with local governments, per TechCrunchReportedTechCrunchOct 10Policy & safety
- Nikon disqualifies the winner of its Small World in Motion video contest for using generative AI, per The VergeReportedThe VergeOct 10Policy & safety
- Seattle bars grocery stores from setting prices with AI based on personal data, a post citing the mayor's office saysUnconfirmedr/singularityOct 9Policy & safety
- Arena launches the Arena Alignment Index, scoring 27 models on real-world agent safety
- Teen airlifted off Crown Mountain after Claude-guided hike route, per The DecoderReportedThe DecoderOct 9Policy & safety
- Anthropic usage policy now bans sustained abuse of Claude and expands weapons and surveillance rulesOfficialAnthropicOct 9Policy & safety
- Goodfire ships inside-model probes that monitor AI agents cheaper than second-model watchers, per TechCrunch
- Three fired OpenAI safety researchers dispute misconduct claims in an open letter, per TechCrunchReportedTechCrunchOct 9Policy & safety
- OpenAI says it disrupted two AI-enabled false-front influence operations using fake journalists and a think tank