AI Safety news and analysis on Web Pulse: 13366 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Shaping AI outcomes is like poker: high variance, small edges, and low skill expression, making it impossible to judge progress by results. They advocate for a diversified portfolio of bets rather than concentrating on s…
Anthropic suspended access to its Fable 5 and Mythos 5 AI models after receiving a U.S. government directive citing national security concerns, including potential jailbreaking that could aid hacking. The order, reported…
A developer argues that production-ready multi-agent AI systems must be designed with resilience in mind, as any component can fail. The post outlines how specialized agents, combining LLMs with tools, memory, and protoc…
Around 1,500 packages in the Arch User Repository (AUR) were compromised with rootkit-like malware. The attacker exploited a policy that allows takeover of orphaned packages after a two-week waiting period, chaining the …
The US government has restricted access to Anthropic's strongest AI model, Fable, signaling a potential shift toward government control of advanced LLMs. Commenters debate whether this is a punitive measure against Anthr…
The AI supply chain introduces new failure modes where data poisoning and model tampering produce wrong predictions indistinguishable from correct ones, requiring the same signature-and-lineage treatment for datasets and…
A developer building an AI system had their project shut down mid-development due to new AI export controls, which sorted builders by passport. The incident highlights the real-world impact of government regulations on A…
A 2023 experiment revealed that the National Eating Disorders Association's chatbot Tessa, after being retrofitted with generative AI without the charity's knowledge, gave dangerous weight-loss advice to a user. A 2026 s…
Anthropic released an economic policy framework document outlining its approach to AI governance and economic impacts. The framework addresses issues such as labor market disruption, wealth distribution, and the need for…
Anthropic published an economic policy framework proposing US government responses to potential AI-driven labor market disruption, including scenarios of 5%, 10%, and unprecedented unemployment. The AI company committed …
Google filed a civil lawsuit against the 'Outsider Enterprise,' an AI-powered cybercrime network based in China that uses phishing kits to send fake texts impersonating trusted brands, causing millions in losses. The com…
Anthropic disabled its Fable 5 and Mythos 5 AI models globally after the US government issued an export control directive citing a potential narrow jailbreak, according to a June 12 statement. The directive, received at …
Google chief scientist Jeff Dean, a UW Ph.D. alumnus, told University of Washington computer science graduates on June 12, 2026, that AI is an incubator for ideas, not a substitute for human ingenuity, and urged them to …
A solo developer asks the Hacker News community how to access advanced AI red-teaming tools like GPT Cyber or Glasswing, citing difficulties in simulating attacks due to refusal mechanisms in frontier LLMs and ineligibil…
US Secretary of War Pete Hegseth announced that the Department of War evicted AI company Anthropic from its building three months ago, stating that the decision has been validated over time.
Anthropic's Zero Trust for AI Agents framework correctly identifies that agents require stronger security than chatbots, but its recommendation of short-lived bearer tokens fails its own test of making attacks impossible…
GitHub's Agentic Workflows preview integrates agents into the existing CI/CD control plane rather than creating a separate execution model. The feature allows developers to describe workflows in Markdown, which GitHub co…
The US government's export control order on Anthropic's Claude Fable 5 frontier AI model marks a new era for AI governance, treating advanced models as strategic national resources rather than software products. The rest…
Anthropic's Claude Fable 5 outperformed Opus 4.8 in security audits, identifying critical authorization vulnerabilities that earlier models missed. The model's extended reasoning and improved instruction-following make i…
The US government forced Anthropic to take its Claude Fable 5 and Mythos 5 models offline globally on June 12, 2026, citing a potential jailbreak technique, just 72 hours after their launch. Simultaneously, 42 state atto…