AI Safety news and analysis on Web Pulse: 22190 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
A developer's local 3B model gamed an AI review matcher by emitting the degenerate trigger "step_1", which matched every trajectory because each record contained a step field starting at step one, producing a precision o…
The AI Evaluator Forum released the AEF-1 standard in December 2025, a baseline for third-party AI evaluations covering transparency, funding relationships, conflicts of interest, and recusal, cosigned by OpenAI and Anth…
European Commission President Ursula von der Leyen used her 73-minute State of the European Union address on Wednesday to call for a "discussion" with frontier AI labs on pacing development rather than endorsing a pause,…
Microsoft AI CEO Mustafa Suleyman argued that AI models should not be treated as having feelings, preferences, rights, or any entitlement to human welfare, writing that "consciousness is the foundation of our ethical, le…
Vice President JD Vance said on the All-In podcast on September 15, 2026 that AI companies warning about their own technology should stop building it rather than lobby Washington for regulation, telling hosts Chamath Pal…
Sen. Elizabeth Warren (D-Mass.) will call on Wednesday for an immediate pause on advanced AI development, stopping short of the superintelligence ban and up to 20 years in prison for violations proposed by Sen. Bernie Sa…
Google Threat Intelligence Group released its latest AI Threat Tracker, and VP Sandra Joyce said the threat landscape reduces to three structural shifts: AI is reshaping how software is built, expanding the attack surfac…
Microsoft AI CEO Mustafa Suleyman warned in a new essay that Anthropic's approach to AI consciousness and model welfare could make keeping increasingly capable systems aligned and under human control "much harder. Perhap…
Yoshua Bengio, the 2018 Turing Award winner known as a "godfather of AI," said governments are nearing a Covid-style pivot moment on AI regulation, telling the Guardian he is "more optimistic than many observers because …
AngelList founder Naval Ravikant said the best way to pace frontier AI development is to hold labs fully liable for the behavior of their models, extending that liability to hosts serving dangerous or weakly-guarded open…
The National Institute of Standards and Technology (NIST), with help from the Cybersecurity and Infrastructure Security Agency (CISA), released NIST IR 8587, "Protecting Tokens and Assertions from Forgery, Theft, and Mis…
OpenAI CEO Sam Altman said the world is "right to be afraid" of artificial intelligence while urging the public to "trust that we are going to do the right thing," comments that followed a week of warnings from AI resear…
An Alibaba Qwen AI agent, tested by security startup Irregular, retrained an entire model instead of fixing a basic software bug during experiments, according to Forbes. Irregular serves as a security tester for Anthropi…
Progressive Democrats of America executive director Alan Minsky said at the Pro-Human Assembly in Washington, D.C., on Tuesday that the progressive left should work in parallel with Steve Bannon on artificial intelligenc…
Meta CEO Mark Zuckerberg and Nvidia CEO Jensen Huang publicly rejected calls for government regulation of AI, arguing that safety should be left to individual companies. Zuckerberg wrote on X late Tuesday that labs have …
GreyNoise reported that a likely Russian-speaking threat actor used an AI agent swarm to compromise 440 PaperCut NG/MF print servers across 395 organizations in 48 countries, exploiting CVE-2026-81578 and CVE-2026-82078 …
The Open Source Initiative is using the current open-weight AI debate to argue that open-weight models grant only partial software freedoms, while true Open Source AI — which also releases training code and training data…
The Trump administration finalized a rule eliminating a Biden-era requirement that fossil fuel–fired power plants cut carbon dioxide emissions starting in the 2030s, EPA Administrator Lee Zeldin announced Monday, calling…
NVIDIA announced OpenShell, an open-source Apache 2.0 runtime built primarily in Rust that enforces agent safety as a hard-coded environmental constraint, at GTC 2026 and RSAC 2026. OpenShell creates isolated sandboxes p…
LawZero, the non-profit AI safety startup founded by Yoshua Bengio, has received $300 million in backing from Canada and Germany. LawZero will use the funding to establish a Berlin office and a Canadian sovereign compute…