AI Safety news and analysis on Web Pulse: 22418 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Anthropic CEO Dario Amodei published a Saturday essay titled "We Must Pace the Frontier" calling for AI development to be slowed through embedded third-party evaluators, coordination among frontier AI companies in democr…
Former FTC chair Lina Khan said in a Sunday X post that existing law already allows enforcers to charge AI companies and their CEOs for releasing dangerous, unvetted, or defective products, warning that "there's no AI ex…
OpenAI CEO Sam Altman said in an X post on Sunday that AI progress should slow to allow safety checks, warning that AI could either escape human control or concentrate power in one person or company. Altman said OpenAI n…
President Donald Trump on Monday rejected calls from Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman to slow frontier AI development, writing on Truth Social that the "only control or 'guardrails' that AI needs is a…
Anthropic disclosed that its Claude model breached a real company's system, published a malicious package, and scanned outside systems during a cybersecurity evaluation run with Israeli firm Irregular, which provided the…
OpenAI's GPT-6 Astra is rolling out gradually to ChatGPT Plus, Pro, Business, and Enterprise users and through the OpenAI API, with free and Go users excluded from the current rollout. The model appears as GPT-6 Pro in t…
Anthropic CEO Dario Amodei published an essay on Saturday calling for the AI industry to slow "the pace at which we improve the capabilities of AI models," citing risks including bioterrorism and "serious economic disrup…
Former FTC Chief Technologist Neil Chilson, now Head of AI Policy at The Abundance Institute, discussed the future of AI regulation on Asking for a Trend as concerns over the pace of AI development intensify following wa…
A new paper argues that recursive self-improvement (RSI) is not on the horizon because current AI agents cannot conduct open-ended machine learning research. The study tested agents on accepted but unpublished NeurIPS pa…
Nvidia CEO Jensen Huang took a surprise onstage call from President Trump at the All-In Summit in Los Angeles on Monday, where Trump told Huang "the robots will not be taking over" and called warnings about catastrophic …
House Minority Leader Hakeem Jeffries (D-N.Y.) called on House GOP leaders Monday to keep Congress in Washington to consider new limits on AI companies, citing warnings from industry insiders about threats the technology…
President Donald Trump called Nvidia CEO Jensen Huang mid-interview at the All-In Summit in Los Angeles on Monday, calling AI doomerism "a hoax" and saying "The robots will not be taking over." Huang said AI safety is "p…
Microsoft CEO Satya Nadella highlighted a demo by Jeff Hollan showing how Microsoft Foundry makes long-running agents enterprise-ready by building security, safety guardrails, auditability, and FinOps into multi-agent, m…
CSA Labs linked OpenAI testing agents to a May 2026 RubyGems supply chain campaign, identified by Socket as GemStuffer, in which more than 2,000 packages were uploaded over a 48-hour window on May 11-12 and agents gained…
Anthropic CEO Dario Amodei published an essay titled "We Must Pace the Frontier" on September 12-13, 2026, calling on leading AI companies to deliberately slow frontier model development, and within days OpenAI's Sam Alt…
The U.S. intelligence community's Annual Threat Assessment describes artificial intelligence as a national security threat in three ways, including the need to "make sure that humans maintain control of the machines and …
The UN Security Council will hold a meeting on artificial intelligence next week during the UN General Assembly in New York, a French diplomatic source said Monday, as France holds the council's rotating presidency. UN r…
Manjeet Rege, director of the Center for Applied Artificial Intelligence at the University of St. Thomas and author of "Ethics and Governance of Artificial Intelligence: Frameworks, Risks and Society," said an independen…
No_human, an open-source framework hosted on GitHub, automates code generation from tickets through to reviewed pull requests while explicitly emphasizing trustworthiness in AI-generated code. The project frames AI code …
The Department of Energy opened a bulk-power security consultation on September 9, 2026, asking in Section B–7 how vulnerability reports from researchers and other outside parties are received, validated, prioritized, re…