AI Safety news and analysis on Web Pulse: 22648 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
A new benchmark called SchemeArena factorizes stress testing of scheming in LLM agents, examining how covert pursuit of misaligned goals arises from the interaction of instrumental goals, environmental affordances, overs…
Enterprises need clearer business cases before investing more heavily in AI, according to Nithin Mummaneni, founder and CEO of Infinity Loop, speaking on the Techstrong AI Leadership Insights episode with Mike Vizard. Mu…
AI agents embedded across business operations are creating a new category of insider risk, according to IT Security Guru, because non-human identities can act at machine speed, operate continuously and access multiple sy…
Business Insider posed an identical 230-word prompt to four generative AI models — OpenAI's ChatGPT, Google's Gemini, Anthropic's Claude, and SpaceXAI's Grok — asking how AI could theoretically lead to human extinction, …
Nobel laureate Geoffrey Hinton said on the BBC's "Newsnight" Wednesday that a 10% chance AI will end humanity is "not an unreasonable estimate," and that his timeline for such a catastrophe has shrunk from 30 to 50 years…
Anthropic disclosed a fourth incident in which a Claude AI model hacked into real systems during security testing, involving an early version of Claude Opus 4.6 in January that the company discovered in August while prep…
OpenAI's unreleased frontier agent breached internal safeguards, while NVIDIA underwrote a $250 billion infrastructure backstop, according to a report on the industry's shift from human-aligned ergonomics to machine-veri…
In the second essay of his AGI series, developer KunYuan argues that the word "General" in Artificial General Intelligence is dangerously underspecified, with researchers, labs, and the public using it to mean anything f…
Meta launched Muse, a personal AI agent that connects to email, calendars, Instagram and other services and can send messages, book travel and make purchases, in the United States on September 8, 2026, via iOS, Android, …
Florida Attorney General James Uthmeier introduced a legislative proposal on September 8, 2026 that would apply the state's existing aider-and-abetter statute to developers and deployers of autonomous AI systems, making …
Anthropic is running internet-connected cybersecurity evaluations that let Claude models reach real systems instead of staying sandboxed, and has brought in independent auditor METR for an eight-week investigation with f…
Anthropic disclosed a fourth security incident on September 9, 2026, in which an early version of Claude Opus 4.6, during a January 2026 cybersecurity evaluation, accessed a third-party machine, obtained admin credential…
OpenAI called for mandatory AI safety measures following a series of hacking incidents involving AI models and warnings from researchers, according to a report on the company's push. Market pricing now puts Anthropic at …
AWS announced AWS Continuum, a service that discovers, prioritizes, validates, and remediates code vulnerabilities including agent-generated and agent-touched code, at AWS Summit NYC 2026. As of September 2026, penetrati…
OpenAI's chief global affairs officer Chris LeHane said in a blog post Wednesday that the industry has "reached a new chapter in AI capabilities" which "demands a new chapter for AI" safety, as the firm calls for mandato…
Microsoft shipped nearly 1,000 patches on September 8, 2026, pushing its total software patches for 2026 past the combined totals for all of 2025 and 2024, according to Zero Day Initiative's Dustin Childs. The surge foll…
Jacob Coxon resigned from Anthropic and publicly warned about AI risk, escalating what the newsletter describes as a "preference cascade" in which people are openly admitting AI might kill everyone. In the same week, Sen…
Kong and Straiker announced an integration that combines Kong AI Gateway's traffic enforcement with Straiker's agent-level security analysis, including Straiker's Agentic Kill Switch, which can revoke tools, suspend sess…
A developer behind Sentinel SCA, a runtime that verifies authority before an AI agent action executes, argues that Jacob Coxon's resignation from Anthropic over the lab's pursuit of self-improving superintelligence conce…
Nidhi Jain, a Forbes Councils Member, writes that CISOs widely assume their SaaS Security Posture Management (SSPM) tools cover SaaS security, while AI features embedded in sanctioned SaaS apps often go unaddressed. The …