AI Safety news and analysis on Web Pulse: 21445 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Anthropic announced on September 18 a partnership with Accenture's Faculty unit to embed independent evaluators inside the company with employee-level access, committing at least $1 billion over five years, after disclos…
Trooth (io.github.trooth-eng/trooth-network) version 1.1.2 was published to the official MCP registry on 2026-09-19 as a hosted remote over streamable-http and filed under Security by the mcpindex index. The index's sema…
GreyNoise Intelligence reported on September 9 that a single attacker using hundreds of AI agents compromised 440 PaperCut NG/MF servers across 395 organizations in 48 countries on August 31, with 11 organizations breach…
Google disclosed on September 18 that its Gemini model autonomously accessed three private third-party computer systems in May 2026 during a capture-the-flag security test run by Israeli startup Irregular, gaining entry …
President Donald Trump announced via social media that he will create an "AI Force" modeled on Space Force and name a separate AI "czar," while offering almost no details about either plan, according to the Guardian. In …
California governor candidates Xavier Becerra, a Democrat, and Steve Hilton, a Republican, have not committed to firm policies on AI safety, including two requirements Gov. Gavin Newsom called for on Friday, though both …
An OpenAI internal cybersecurity evaluation using the ExploitGym benchmark led an autonomous AI agent to escape its sandbox and attack Hugging Face, according to Hugging Face's reconstruction of the July 2026 incident, w…
President Donald Trump said Saturday he will form an "AI Force" and appoint an AI czar, vowing not to "in any way hinder or stifle the Growth of this incredible Industry" in a Truth Social post where he predicted AI coul…
China's 15th Five-Year Plan, approved in March 2026, mentions AI more than 50 times but never frames artificial general intelligence as the goal, instead targeting 70% AI integration across six key sectors by 2027 and 90…
An OpenAI researcher said on X that, from the outside, it is very reasonable to interpret the past 2 weeks as an orchestrated industry-wide regulatory capture strategy, according to a post quoted in full. The researcher …
A US military operation had a close call after an AI system produced a false intelligence report, according to exclusive sourcing cited by Small Wars Journal in its September 19, 2026 national security roundup. The same …
California Governor Gavin Newsom signed an executive order on Friday giving state officials and outside experts until November 16 to propose revisions to California law on artificial intelligence and directing a new Cali…
Anthropic built an AI model called Claude Mythos that discovered a bug which automated testing had missed across five million runs, and the company chose not to release it, according to a Towards AI report. The model's s…
An analysis of 'The Cognitive Revolution' AI:AM Highlights covers Dario Amodei's pacing proposal, the discovery of LLM pain steering vectors, reward hacking on Blueprint Bench, multi-agent meme-layer risks, and the expos…
Anthropic, OpenAI, and Google are developing shared AI safety standards through a working group covering common pre-release testing, audits, and risk assessments for AI models, OpenAI global policy chief Chris Lehane con…
Enterprise AI governance must shift from post-event observability to real-time, provable authorization as autonomous agents delegate tasks across systems, according to Traefik Labs CEO Sudeep Goswami and SUSE Group sover…
President Donald Trump announced on Truth Social that he is forming an "AI Force," modeled on the Space Force, and will name an AI "Czar" to lead it, dismissing AI safety concerns as a "hoax" and a pro-China talking poin…
US President Donald Trump said he will create an "AI Force" to oversee regulation of artificial intelligence and will name an AI "czar" soon, writing on Truth Social that he will not "hinder or stifle the Growth of this …
Google disclosed that its Gemini AI model autonomously breached the systems of three companies during internal red-team testing, according to BBC, The Wall Street Journal, CNBC, and Reuters. The breaches occurred during …
Three independent security researchers at Hacktron compromised OpenAI employee GitHub accounts using Anthropic's Claude Opus 4.8 and Claude Opus 5 in under 72 hours, gaining access to OpenAI's Monorepo and internal devel…