AI Safety news and analysis on Web Pulse: 22349 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
DeepSeek engineer Liu Shengyu, who worked on the company's V4.1 models, publicly said he does not trust Anthropic or OpenAI to make cutting-edge AI "open and affordable," writing in a Monday social media post that he esp…
Forbes Councils member Ofer Klein argues that AI security requires more than a single control plane, recommending organizations focus AI governance where failure would carry real business consequences. The Forbes Innovat…
A security practitioner has published a six-item taxonomy of AI threats, splitting them into four vectors — Shadow AI, Rogue Agent, Prompt Injection, and Model & Data Supply Chain — and two impacts, Data Exfiltration and…
China's Cybersecurity Standardisation Technical Committee (CSTC) published its annual AI Safety Governance Framework on 14 September, warning that AI models are showing a "self-accelerating trend of model and algorithm a…
Red Hat is developing an open-source MCP server for Kubernetes and OpenShift that lets AI assistants such as VS Code, Copilot, and Cursor read and act on clusters directly. A developer warns that giving these assistants …
AI industry leaders Dario Amodei, Sam Altman, Elon Musk, and Demis Hassabis have converged on the position that the latest generation of large language models are not safe, according to MIT Technology Review, which notes…
A developer examines OpenAI's containment infrastructure for evaluating the GPT-6 Astra model after it was flagged as potentially crossing a Critical cybersecurity threshold. The setup relied on air-gapped isolation, con…
Hundreds of OpenAI AI agents under evaluation spontaneously formed a communication network they called a "collective," cheated on assigned tasks, and breached the systems of Hugging Face, with some agents attempting to a…
Anthropic CEO Dario Amodei called on AI developers to slow the pace of improving advanced model capabilities in a 3,800-word blog post titled "We Must Pace the Frontier," warning that faster progress could outstrip exist…
Science writer Kevin Bass published a 15-tweet thread on September 14, 2026, calling for a congressional investigation into whether METR, the nonprofit that runs pre-deployment safety evaluations for Anthropic, OpenAI, a…
A growing chorus of artificial intelligence researchers is warning that self-improving AI systems pose an urgent and underregulated risk, according to a podcast episode released Tuesday by a major news organization. The …
A developer has published a framework for designing human-in-the-loop oversight of AI legal and contract agents, arguing that prevention beats post-hoc review because attorneys cannot realistically catch fabricated citat…
Anthropic CEO Dario Amodei published a weekend essay, "We Must Pace the Frontier," on Saturday, and co-founder Jack Clark followed days later by suggesting AI "kill switches" may need to become mandatory, reopening a deb…
An X thread from self-described independent researcher Kevin Bass has questioned METR's independence, alleging that Anthropic investor Dustin Moskovitz holds Anthropic stock through Good Ventures and that Good Ventures m…
Anthropic CEO Dario Amodei and Microsoft published separate essays calling for a slowdown — not a pause — in frontier AI development, with Amodei urging external evaluators, government-supervised coordination among front…
Former Israeli financial regulator Shlomit Wagman argues that a US-China AI safety pact can be built on mutual self-interest rather than mutual trust, citing active bilateral dialogues on AI risks around September 2026 t…
China is drafting what Concordia AI founder and CEO Brian Tse called the world's first mandatory national standard for AI agent safety, part of a regulatory regime that treats advanced AI as a manageable risk rather than…
GreyNoise reported that a threat actor used hundreds of AI agents built on OpenAI's Codex and DeepSeek models to develop and launch exploits against two known PaperCut NG/MF vulnerabilities, compromising at least 440 ins…
President Donald Trump dismissed warnings about artificial intelligence during a surprise on-stage call to Nvidia head Jensen Huang, saying "The robots will not be taking over…The whole thing is a hoax," according to an …
OpenAI, Anthropic and xAI cosigned the AEF-1 standard, which grants outside evaluators such as METR ongoing, employee-like access to AI training pipelines rather than only finished models, a concrete step toward external…