AI Safety news and analysis on Web Pulse: 22439 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
A security engineer built m8m, a tool for auditing the persistent memory files that AI coding agents like Claude Code, Codex, and Cursor write automatically, after finding no way to inspect what those agents stored or wh…
President Donald Trump attacked Anthropic CEO Dario Amodei on Truth Social on Monday after Amodei published a Saturday essay calling for frontier AI labs to "slow the pace at which we improve the capabilities of AI model…
Microsoft AI published a code of conduct for its MAI models that prioritizes human control over autonomy and performance, with AI chief Mustafa Suleyman stating, "If it isn't safe we shouldn't build it." The rulebook rej…
Technology leaders worldwide spent the weekend debating whether leading AI labs should slow development to ensure increasingly capable AI models are safe to use, a debate that is punishing AI stocks, according to the new…
Microsoft released a draft of its "Humanist AI" code of conduct on Monday, stating that AI models should remain under human control and that any models that cannot be contained must be rejected. The preliminary guideline…
Anthropic CEO Dario Amodei published an essay on September 12 titled "We Must Pace the Frontier," urging the AI industry to slow development of advanced capabilities while Anthropic maintains a $517 billion compute commi…
Adversarial fashion uses specific patterns and silhouettes to lower the confidence scores of object detection models or force them to misclassify humans as animals or objects, according to reporting on the practice. In 2…
Anthropic CEO Dario Amodei called for slowing AI capability development, and OpenAI CEO Sam Altman and xAI CEO Elon Musk were among the competitors who followed with similar calls after a researcher warned AI could soon …
Yahoo Finance Tech Editor Dan Howley examined what safety warnings from Anthropic and OpenAI about pacing AI development mean for future regulation, in a video published Monday, September 14, 2026. The segment weighed wh…
OpenAI CEO Sam Altman warned on X of a potential AI control crisis, describing the ultimate AI safety risk as losing control to the machines. Altman, who leads OpenAI, framed the loss of human control over AI systems as …
President Donald Trump said Monday that he would serve as sufficient oversight for artificial intelligence, calling efforts to limit the technology a "SICK conspiracy" and pushing back on calls for government guardrails …
The Midas Project, a nonprofit AI watchdog, alleges OpenAI violated California's Transparency in Frontier AI Act (SB 53) at least three times this year by failing to publish required risk assessments for its GPT-5.6 prev…
President Donald Trump wrote on social media Monday that the only "guardrails" AI needs is a "strong and smart (High IQ!) president," and named Anthropic CEO Dario Amodei as an AI alarmist "pretending to be a 'perfect li…
President Donald Trump rejected new AI guardrails in a Truth Social post, targeting Anthropic CEO Dario Amodei and claiming the U.S. government already holds "tremendous CRIMINAL and REGULATORY power" over AI companies. …
A healthy LLM application must consistently generate useful results while staying within acceptable limits for accuracy, safety, speed, reliability, and cost, according to a technical explainer on LLM evaluation. The pie…
President Donald Trump said in a Truth Social post Monday that artificial intelligence needs no guardrails beyond "a STRONG AND SMART (High IQ!) PRESIDENT," and warned that "we already have tremendous CRIMINAL and REGULA…
OpenAI CEO Sam Altman warned in an X post early Monday that AI progress could go "very badly" in two ways — humanity losing control of the future to AI, or an extraordinarily powerful AI concentrating power in one person…
Anthropic is unilaterally committing to embed third-party evaluators, such as METR, inside its offices with desks, access badges, and company laptops, as the first of three proposals to pace frontier AI capability growth…
Emacs upstream maintainer Sean Whitton announced that the original fix for the arbitrary code execution flaw CVE-2024-53920 was incomplete, after Bas Alberts found that viewing or editing untrusted files in modes other t…
President Trump dismissed AI safety concerns in a Truth Social post and argued the United States must prioritize beating China in AI advancement, responding to Anthropic CEO Dario Amodei's weekend essay calling for slowe…