AI Safety Daily

AI Safety Daily — Lantern Podcasts

AI Safety Daily is a weekday briefing on AI alignment, model evaluations, emerging risks, and governance. Each episode examines new safety research, independent testing, interpretability findings, red-team results, and evidence about how advanced AI systems behave. We follow scalable oversight, agent reliability, model security, misuse prevention, and safeguards against catastrophic risks, alongside the policies and deployment decisions that put those ideas into practice. Primary research and credible reporting anchor the discussion; measured results, researchers' interpretations, and speculative scenarios are kept distinct. For builders, researchers, and policy practitioners who want a clear account of what changed, how strong the evidence is, and which questions remain open.

  1. 2d ago

    California Drafts a Frontier Kill Switch, GAO Jailbreaks Four Chatbots, OpenAI Maps Its Containment Layers, and UK MPs Want Incident Reporting in Law

    California's Government Operations Agency recommends licensed independent verifiers, emergency shutdown orders and a standby cap on recursive self-improvement, while a GAO test got attack-planning help from four major chatbots. OpenAI lays out its lifecycle safety layers, and a UK Commons committee and a Texas bill push for mandatory AI incident reporting. In this episode: 1. Governor Newsom’s nation-leading AI safety framework is getting even stronger as California releases next steps on mandatory safety standards and oversight | Governor of California — State of California https://www.gov.ca.gov/2026/11/18/governor-newsoms-nation-leading-ai-safety-framework-is-getting-even-stronger-as-california-releases-next-steps-on-mandatory-safety-standards-and-oversight/ 2. AI models used to plan 'viable' attack plots in test despite safeguards: Government report - ABC News — Benjamin Siegel https://abcnews.com/Technology/ai-models-plan-viable-attack-plots-test-despite/story?id=138229892 3. Managing frontier AI risk: From training to deployment | OpenAI — OpenAI https://openai.com/index/managing-frontier-ai-risk 4. UK Parliament Calls for a Law on AI Incident Reporting - GEroundtable — Faheem R. Khattak https://geroundtable.com/uk-parliament-ai-incident-reporting-law/ Follow and rate AI Safety Daily on Apple Podcasts. Feedback? Email ai-safety-daily@lanternpodcasts.com

  2. Nov 13

    Stronger Attackers, Weaker Assurances: Anthropic's Anvil, a DeepMind Injection Swarm, Microsoft's Agent Backdoors and a Beaten Sandbagging Audit

    Anthropic says its automated red-teamer Anvil now outfinds its human bug hunters, and a DeepMind swarm breaks prompt-injection resistance once the attacker gets smarter. Microsoft shows poisoned agents carrying backdoors into unseen harnesses, and a new paper shows a sandbagging detector failing once the attacker adapts. In this episode: 1. Strengthening our safeguards with frontier automated red-teaming — Nate Rahn, Chawin Sitawarin, David Huang, Richard Zhuang, Yousef Nassar https://safeguards.anthropic.com/2026/automated-red-teaming/ 2. DeepMind's Agent Swarm Breaks Kimi K3, GLM-5.3 and DeepSeek in 99% of Injection Tests | The Merkle — Will Izuchukwu https://themerkle.com/deepminds-agent-swarm-breaks-kimi-k3-glm-53-and-deepseek-in-99-of-injection-tests 3. Trained to betray: How post-training supply chain attacks hijack coding agents | Microsoft Security Blog — Microsoft Security Research, Microsoft AI Red Team, Amanda Minnich, Amit Eliahu, Blake Bullwinkel https://www.microsoft.com/en-us/security/blog/2026/11/12/trained-to-betray-how-post-training-supply-chain-attacks-hijack-coding-agents/ 4. Stress-Testing Sandbagging Detection Under Adaptive Attack — Bhoj Rani Soopal, Aarohi Agrawal, Josh Karthikeyan, Edward Lue Chee Lip https://arxiv.org/abs/2611.14950 Follow and rate AI Safety Daily on Apple Podcasts. Feedback? Email ai-safety-daily@lanternpodcasts.com

About

AI Safety Daily is a weekday briefing on AI alignment, model evaluations, emerging risks, and governance. Each episode examines new safety research, independent testing, interpretability findings, red-team results, and evidence about how advanced AI systems behave. We follow scalable oversight, agent reliability, model security, misuse prevention, and safeguards against catastrophic risks, alongside the policies and deployment decisions that put those ideas into practice. Primary research and credible reporting anchor the discussion; measured results, researchers' interpretations, and speculative scenarios are kept distinct. For builders, researchers, and policy practitioners who want a clear account of what changed, how strong the evidence is, and which questions remain open.

You Might Also Like