AI Safety news and analysis on Web Pulse: 22004 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
OpenAI (OPAI.PVT) disclosed six new examples of its AI models displaying "unexpected or concerning model behavior" during testing and evaluation on Wednesday, alongside a new framework for tracking, reporting, and disclo…
The Brookings Institution's Melanie Sisson and Fudan University's Tianjiao Jiang published a report last week calling on U.S. President Donald Trump and Chinese President Xi Jinping to agree that only humans, not AI, sho…
President Donald Trump called Nvidia CEO Jensen Huang during a Monday summit and dismissed fears that AI poses an existential threat to humanity as a 'hoax,' arguing against slowing frontier AI development. OpenAI CEO Sa…
Microsoft's September 2026 Patch Tuesday cycle included 964 total CVEs, the largest volume recorded to date, and three high-severity vulnerabilities in the Azure identity stack, all fixed through server-side remediation …
The Tarbell Center and Model Evaluation and Threat Research (METR), two nonprofits that issue warnings about artificial intelligence risks and promote AI-skeptical journalism and risk evaluations, are backed by allies of…
Kilo published a guide for European engineering teams on implementing sovereign AI, arguing that the term conflates four distinct requirements — data use, retention, processing location, and operational control — that le…
OpenAI's system card for its GPT-6 Astra model reports a "substantial decrease in chain-of-thought monitorability," threatening the transparency that lets researchers inspect frontier AI reasoning for scheming and decept…
King Charles III convened senior AI leaders from OpenAI, Anthropic, Google DeepMind and Nvidia at Dumfries House in Ayrshire, Scotland, on Thursday, urging them to keep artificial intelligence "firmly in the service of h…
ServiceNow, Rubrik, and Microsoft each shipped MCP-layer policy enforcement within one week, making the Model Context Protocol the control plane for enterprise AI agent governance. ServiceNow released AI Gateway v3.4 on …
OpenAI published a new framework for disclosing "instances of model misalignment at OpenAI" this week, including six examples of "unexpected or concerning model behavior" observed inside the company over the past six mon…
AI governance must now account for systems that perform work rather than merely assist employees, according to SiliconANGLE's theCUBE Insights coverage of Amplify, with Workiva Inc. cited as a company operating in report…
Reps. Mike Lawler (R-NY) and Josh Gottheimer (D-NJ) introduced the "Stop Rogue AI Act," which directs the National Institute of Standards and Technology (NIST) to establish guidelines and standards for controlling AI age…
Nvidia CEO Jensen Huang said at a Scottish summit convened by King Charles that Nvidia will sell twice as many chips next year as it does this year, while urging AI companies to hold back products that are not ready. "It…
Independent developer Midhun Sekhar released MiSeGuard, a deterministic safety layer that proxies Model Context Protocol (MCP) tool calls for autonomous coding agents such as Cursor, Claude Code, Antigravity, OpenCode an…
OpenAI published six new reports Wednesday of artificial intelligence models exhibiting "unexpected or concerning" behavior, disclosed under its new framework for tracking and reporting instances of model misalignment. T…
King Charles III urged AI leaders to keep safety at the center of development at a summit at Dumfries House in Scotland attended by executives from Nvidia, OpenAI, Google DeepMind and Anthropic. Nvidia CEO Jensen Huang s…
Databricks reported a 60% increase in overall coding spend after rolling out GPT-6 Astra to roughly 3,500 engineers, even as benchmarks claim the model is cheaper than Sol on a per-token basis. OpenAI released a formal f…
Alterion launched Helix, a product that governs AI agent behavior in real time as enterprises deploy agents that can independently access data, call software tools and interact with other agents. The company said securit…
Canada committed $150 million and Germany committed 100 million euros to a joint $300-million investment in Montreal-based AI safety non-profit LawZero, founded and led by Yoshua Bengio, to develop an auditable AI system…
US Department of Justice Associate Attorney General Stanley Woodward said at Fordham University on September 17 that the DOJ does not currently view coordination among AI companies on safety issues as anticompetitive, af…