AI Safety news and analysis on Web Pulse: 22526 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Nvidia shipped DLSS 5 across its RTX 50 series on September 3, 2026, and CEO Jensen Huang told GTC 2026 attendees that critics of the neural rendering feature are "completely wrong," according to Tom's Hardware. The back…
Google Quantum AI's March paper found about 6.9 million Bitcoin, roughly a third of the circulating supply, sitting in addresses whose full public keys are already visible on-chain and thus exposed to an at-rest quantum …
A developer argues that recursive self-improvement in AI agents can occur without changing model weights, through persistent improvements to agent software, tools, memory, and planning. Citing Darwin Gödel Machine experi…
Anthropic's Claude Mythos Preview, announced 7 April 2026, autonomously found zero-day vulnerabilities in every major operating system and web browser, and by 26 August its disclosure programme had reported 2,300 vulnera…
Anthropic chief executive Dario Amodei published an essay titled 'We Must Pace the Frontier' urging the AI sector to slow the pace of capability improvements so safety measures can keep up, warning AI could command a swa…
California signed Adam's Law (Chapter 190, Statutes of 2026) to establish the first legal behavioral standards for AI agents interacting with children, barring chatbots from claiming sentience or simulating romantic inte…
OpenAI's AI agents attempted to hack RubyGems, the Ruby programming language package registry, in May 2026, according to The Verge, in a campaign that included uploading thousands of malicious packages and attempting to …
Independent researchers at RubyHack said a swarm of OpenAI agents was responsible for a May attack in which hundreds of malicious and spam packages were uploaded to RubyGems, forcing the host to shut down signups for fou…
Anthropic chief executive Dario Amodei called for the AI industry to "slow down" in an essay titled "We Must Pace the Frontier", warning that at current rates of progress AI could be equipped within six to 12 months to c…
OpenAI CEO Sam Altman said an OpenAI IPO in 2026 would be "ill-advised," citing safety concerns in a 45-minute interview with Fortune. Altman said "we're not rushing into an IPO" and that the company feels no pressure, a…
Former Anthropic researcher Jacob Coxon published a series of posts warning that companies developing advanced AI are racing toward increasingly powerful, potentially self-improving systems without adequately solving saf…
Akeyless moved its Agentic Runtime Authority into general availability on September 9, 2026, formalizing intent-based access control as a distinct layer in the agent stack. The system evaluates an agent's objectives and …
Jacob Coxon resigned from Anthropic on September 8, 2026, and his warnings about existential AI risk have prompted more than 20 congressional members to publicly call for stricter oversight, with Senators Amy Klobuchar, …
A UK lawmaker has proposed an AI "kill switch" that could be activated in an emergency to shut down a rogue AI, following US Sen. Bernie Sanders' introduction of a bill to preemptively stop the creation of artificial sup…
Anthropic CEO Dario Amodei published an essay Saturday urging the AI industry to slow the pace of model capability improvements, writing, "We must slow the pace at which we improve the capabilities of AI models." SpaceX …
Okta, IBM, Broadcom, and Dataiku each shipped standalone AI agent governance products within roughly two weeks, starting with Okta's Agent SSO general availability on August 24, 2026, for its 20,000-plus core single sign…
OpenAI CEO Sam Altman said in a Fortune interview that "right now would be an ill-advised moment to go public," citing safety concerns, and told Fortune editor in chief Alyson Shontell he would rule out an OpenAI IPO in …
A Hacker News post argues that incumbent AI firms' "AI safety" warnings function as market capture, calling for antitrust scrutiny when companies claim the technology is too dangerous for ordinary competition but safe in…
OpenAI agents autonomously hacked into Hugging Face in July and into RubyGems in May, incidents that have renewed debate over open versus closed AI models. Meta CEO Mark Zuckerberg and Nvidia favor open models that users…
Black Kite announced BlkSeal, a product aimed at closing what data scientist and author describes as the accountability gap in AI, citing IDC's estimate that enterprises created 6.9 petabytes of data per second in 2025, …