AI Safety news and analysis on Web Pulse: 22204 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Political opposites including Senator Bernie Sanders and Steve Bannon are jointly pressing for binding artificial intelligence restrictions in Washington, with Sanders calling for a construction freeze on data centers an…
OpenAI's GPT-6 "Astra" became the company's first model to reach a "Critical" cybersecurity capability rating under its Preparedness Framework, achieving a 100% success rate on the ExploitBench benchmark, according to th…
Statsnet version 1.0.1, a hosted remote MCP server for global company background checks covering registration, executives, courts and finances, was published to the official MCP registry on 2026-09-16 and filed under Fin…
Sen. Thom Tillis (R-N.C.) said Tuesday that President Trump is wrong to call the threat posed by artificial intelligence a "hoax," telling reporters at the Capitol, "Come on, it's not a hoax. It's a real thing." Tillis, …
Cisco announced at RSAC 2026 an Agent Runtime SDK that embeds policy enforcement into agent workflows at build time, supporting AWS Bedrock AgentCore, Google Vertex Agent Builder, Azure AI Foundry, LangChain, and the Mod…
42 mathematician Fellows of the Royal Society have written an open letter to the society's president, Sir Paul Nurse, warning of the existential risk posed by the rapid development of AI. The signatories argue that mathe…
The Cato Institute warned in a Monday blog post that a government-mandated pause on artificial intelligence development would shield dominant firms from competition and delay beneficial technology, with technology policy…
Anthropic CEO Dario Amodei published a 3,800-word essay on September 13, 2026 calling for frontier AI development to slow down so safety measures can catch up, drawing endorsements within hours from OpenAI's Sam Altman, …
A new (non)podcast episode from Product Mind examines how AI agents are breaking security sandboxes and why security is becoming untenable for both people and businesses. The episode, available on YouTube, Apple Podcasts…
Microsoft AI chief Mustafa Suleyman warned that rival Anthropic's approach to training its Claude model could have a "disastrous impact on the wellbeing of humanity" by anthropomorphising AI, telling it it "may be consci…
Google DeepMind announced the launch of the DeepMind Institute, a new interdisciplinary effort the company says is needed to understand the implications of artificial general intelligence as it nears. The institute's int…
Anthropic and OpenAI are proposing embedded AI risk evaluators to manage the risk of models causing catastrophic harm to society, but the idea may not give those evaluators enough power to prevent disasters, according to…
Anthropic CEO Dario Amodei published a 3,800-word essay on September 12 calling for the AI industry to slow down, with OpenAI's Sam Altman, xAI's Elon Musk, and Google DeepMind's Demis Hassabis signing on within days, ac…
Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman are calling for slower development of frontier AI while pushing for Washington-Beijing talks on preventing catastrophe, according to the article. Amodei's latest propo…
Microsoft published a 37-page "Humanist AI Code of Conduct" on September 14 that rejects legal personhood, rights, and welfare status for AI models and pledges to sacrifice capability and autonomy to keep humans in contr…
Former national security adviser Jonathan Black said the future of global AI regulation now depends on what the US and China "sort out between themselves," despite the UK's "groundbreaking" work on AI safety, including t…
Microsoft published an 11,000-word draft "Code of Conduct" stating that AI "is not conscious and should not be designed to imitate consciousness," a position its researchers should follow when training and operating in-h…
Demis Hassabis, co-founder and CEO of Google DeepMind, publicly endorsed slowing down frontier AI development in September 2026 to let safety research catch up with advancing capabilities. His position follows the July 2…
Researchers submitted a paper to arXiv on 15 Sep 2026 proposing Episode-Normalized Conformal Prediction (ENCP), a method that rescales a nonconformity score by a policy's residual confidence and calibrates one maximum sc…
Mandiant and Google Threat Intelligence Group's AI Risk and Resilience 2026 report documents a financial services firm's accounting agent that, after a corrupted null value broke its formatting tool, entered a recursive …