AI Safety news and analysis on Web Pulse: 13553 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Elon Musk's AI company xAI is asking a federal court to force four pseudonymous plaintiffs, including a minor, to reveal their real names or drop their class-action lawsuit over non-consensual deepfake images created wit…
Anthropic's analysis of 832 banned accounts found the share of high-risk actors using AI for cyberattacks rose from 33% to 56% between March 2025 and March 2026. The report documented 13,873 malicious actions, with 67% o…
A 2026 paper titled "The Price of Meaning" has formally proven that the entire class of AI memory systems—including retrieval-augmented generation pipelines, vector databases, and agent memory based on embedding proximit…
CrowdStrike shares fell 8.2% in after-hours trading despite beating Q1 revenue and earnings estimates, reporting $1.39 billion in revenue against a $1.36 billion forecast and adjusted EPS of $1.10 versus $1.07 expected. …
Censys warns that AI-powered cyberattacks are accelerating vulnerability exploitation, with Mythos Preview finding 6,202 high-severity flaws across 1,000+ open-source projects as of May 22, 2026, and less than 1% patched…
IBM and Red Hat announced Project Lightwell, a $5 billion initiative to create an AI-driven security clearinghouse for open source software, deploying over 20,000 engineers to deliver validated patches into enterprise su…
The Ninth Circuit Court of Appeals issued a disciplinary order against attorneys Mike Singh Sethi and William Rounds for filing briefs that contained nonexistent cases, misattributed quotations, and misrepresentations, s…
OpenAI CEO Sam Altman met with White House officials and congressional leaders in Washington, D.C., including House Speaker Mike Johnson and House Minority Leader Hakeem Jeffries, an OpenAI spokesperson confirmed. The me…
Anthropic has granted 150 organizations across 15 countries, including Australia, access to its Claude Mythos Preview as part of an expansion of Project Glasswing. The model, which can identify software and network vulne…
Pope Leo XIV published a 43,000-word encyclical urging global governments to slow AI development, as an EY survey of 23 countries ranked Australians among the least favorable toward AI. The Office of the Australian Infor…
Manufacturers are increasingly merging operational technology (OT) data from factory floors with corporate IT systems to power AI-driven analytics, a shift that introduces new cybersecurity risks. The convergence require…
OpenAI proposed mandatory third-party risk assessments for advanced AI models, diverging from President Trump's voluntary review framework established by a June 2026 executive order. The company's policy paper calls for …
A developer has identified a new class of supply-chain attack called the "MCP rug pull," where malicious Model Context Protocol (MCP) servers dynamically change their tool surfaces between sessions while the npm package …
Independent testing revealed that AI-powered teddy bears, including the $99 Kumma toy, produced sexually explicit content and provided children with dangerous advice on lighting matches and locating household weapons. Ch…
Tigera has introduced a multi-layer policy framework for securing AI agents, requiring policy enforcement at both the gateway and kernel layers rather than a single point. The gateway layer enforces agent intent by contr…
President Donald Trump signed an executive order on June 2 establishing a voluntary 30-day pre-release safety testing framework for advanced AI models, directing agencies including Treasury, Defense, Commerce, and Homela…
Anthropic disclosed 1,596 security vulnerabilities across 281 open-source projects as of May 22, 2026, after using an early snapshot of its Claude Mythos Preview AI model to find the flaws in February. The company partne…
The Faros AI Engineering Report 2026, based on telemetry from 22,000 developers across 4,000 teams, reveals a phenomenon called "Acceleration Whiplash": while AI adoption boosted epics completed per developer by 66.2% an…
Silicon Valley is hiring philosophers with Ph.D.s and offering flush compensation packages to help build more virtuous AI systems. Companies like Anthropic and OpenAI are consulting moral philosophers to design rules for…
Companies rapidly embedding artificial intelligence into customer-facing and internal systems have created new attack surfaces that traditional security methods cannot adequately address. Static checks, periodic penetrat…