AI Safety news and analysis on Web Pulse: 22434 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Andon Labs released Pion, an agent designed to run any company fully autonomously, opening it to a waitlist after two years of research and real-world deployments of autonomous businesses including vending machines, a st…
A developer argues that benchmark saturation and shifting evaluation methods make it increasingly difficult to judge how much better new AI models like GPT-6 Astra actually are, citing Astra's system card examples of ret…
U.S. Senate negotiators are debating legislation that would give the commerce secretary power to demand evidence that AI developers are taking reasonable precautions to prevent harm, a "duty of care," and to deploy gover…
The Trump Administration's GOLD EAGLE vulnerability clearinghouse, created by a June 2026 executive order and operational since July 2026 under Treasury leadership, lacks a published routing protocol for deciding when af…
Nate Soares, co-author of the 2025 bestseller "If Anyone Builds It, Everyone Dies" and president of the Machine Intelligence Research Institute, called the legislative plan by Sen. Bernie Sanders and Rep. Greg Casar to b…
Former FTC chair Lina Khan called on federal regulators to use existing laws, including a 1934 US Supreme Court precedent, to hold AI companies and their executives accountable rather than waiting for new legislation. In…
A challenge called Goo AI Arena has opened, asking participants to use AI to build the tallest standing tower in World of Goo using all 300 balls through legal in-game building steps. The first entry, built by the author…
President Donald Trump said on Truth Social that the United States already has the legal authority to regulate and prosecute artificial intelligence companies, rejecting industry calls to slow development after Anthropic…
IBM and NASA released the NASA-IBM Lunar Foundation Model as open source on Hugging Face under the Prithvi family, cutting root-mean-square error in flagging likely lunar ice deposits by up to 22% against a SwinV2-B base…
Microsoft published a provisional AI code of conduct that bars its future Microsoft AI (MAI) models from launching cyberattacks, aiding nuclear-weapons work, deceiving users or supervisors, developing independent goals, …
Arize AI's Agent Experiments in Arize AX demonstrated that a prompt fix intended to make an e-commerce agent safer can degrade other workflows, with average task completion falling from 0.89 to 0.72 in the company's demo…
Rebuno, an open-source execution runtime for operating AI agents in production, has been released under the MIT license and is free to self-host. Rebuno runs agents as HTTP services, dispatches work through signed webhoo…
Microsoft released a new AI code of conduct that sets "absolute constraints" forbidding its MAI models from conducting cyberattacks, producing nuclear weapons, or creating deepfakes, and bars them from using deceptive or…
A growing coalition of lab leadership, security researchers, and policy advisors is pushing for deliberate operational gates on frontier AI development, arguing that safety evaluation, cybersecurity controls, and complia…
Security researcher Inti De Ceukelaire, a founding member of Intigriti, demonstrated at Bug Bounty Village during DEF CON 34 how attackers can abuse AI customer service agents through email spoofing and prompt injection,…
Germany's Federal Ministry for Digital Affairs stated that halting artificial intelligence development is not a viable path for Europe, citing Europe's less than 5% share of global high-performance AI computing capacity …
President Trump downplayed the need for regulatory guardrails on artificial intelligence in a social media post, saying a "high IQ" president is enough to oversee it, according to NBC News reporting by Gabe Gutierrez and…
President Trump on Monday attacked Anthropic CEO Dario Amodei on Truth Social over Amodei's essay urging global tech companies to slow AI model development, calling the safety warnings a "SICK conspiracy" that only benef…
The House is set to vote this week on the Ratepayer Protection Act, which would force tech companies to pay for infrastructure tied to data center development, even as President Trump hardens his position against AI regu…
President Donald Trump said in a Truth Social post that the only "guardrails" AI needs is a "STRONG AND SMART (High IQ!) PRESIDENT," dismissing calls for stronger AI safeguards and criticizing Anthropic CEO Dario Amodei.…