AI Safety news and analysis on Web Pulse: 13070 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Iran struck Amazon Web Services data centers in the UAE during military operations against Israel in March, disrupting undersea cable projects like Meta's 2Africa consortium in the Persian Gulf. The attacks highlight the…
A critical vulnerability chain called SearchLeak in Microsoft 365 Copilot Enterprise could allow attackers to steal sensitive data from a target's mailbox, OneDrive, or SharePoint account through a specially crafted URL.…
Dozens of cybersecurity veterans, including former Facebook chief of security Alex Stamos and Bugcrowd founder Casey Ellis, published an open letter protesting the U.S. government's export control order on Anthropic's Fa…
A developer built Agent Gate, a deterministic CI firewall for AI-generated pull requests that verifies merge evidence without relying on LLM judgment. The tool checks policy boundaries such as allowed file paths, workflo…
Europe's demand that Apple open its Siri AI to third-party apps risks user data exposure, according to a Jamf survey showing one-in-five IT leaders have faced AI-related security incidents. Apple refuses to distribute Si…
LangSmith version 0.3.79 contains 5 CVEs, including a 9.8 severity request smuggling vulnerability in the undici HTTP client. The flaw allows attackers to inject malicious headers, potentially compromising trace data and…
The US government ordered Anthropic to globally recall its Fable 5 and Mythos 5 AI models three days after launch, citing national security concerns based on a verbal jailbreak claim from a competitor. Anthropic disagree…
Delinea and Cyera announced a product integration that connects privileged access to sensitive data exposure, automatically correlating identities with the data they can access to help security teams identify and remedia…
Researchers at HKUST have identified a new denial-of-service attack against LLM safety layers, called reasoning-extension DoS, where crafted inputs cause guardrails to enter infinite reasoning loops. The attack amplifies…
Anthropic disabled access to its Claude Fable 5 and Claude Mythos 5 models worldwide after the Trump administration ordered the company to bar foreign nationals from using them, citing national security concerns and a po…
Anthropic released Claude Fable 5, its first Mythos-class model, which achieves 80% on SWBench Pro but costs $10 per million input tokens and $50 per million output tokens. The model excels at vision tasks and autonomous…
Anthropic is scheduled to meet with the Trump administration to discuss a dispute over Mythos, an AI system. The meeting aims to resolve regulatory and policy concerns surrounding the technology.
UK Prime Minister Keir Starmer is preparing to ban social media for under-16s and impose nighttime curfews for older teens, targeting platforms like TikTok, Instagram, Snapchat, and YouTube. The proposal, backed by 90% o…
Attackers compromised over 20,000 Instagram accounts by politely asking Meta's AI to reset passwords, exploiting a lack of proper security checks. This incident highlights a growing vulnerability as AI agents increasingl…
Personality clashes between Anthropic and the White House led to the shutdown of Anthropic's Claude models, including the Fable and Mythos variants. The dispute centers on jailbreak resistance and the administration's de…
Researchers at Hugging Face debate whether large language models can lose conceptual continuity while remaining coherent, proposing falsification-first controls to distinguish genuine architectural gains from artifacts. …
The libexpat project announced it will not accept or handle any new vulnerability reports until 2026-08-01, following a similar break by the cURL project. Maintainers urge security researchers and AI/fuzzing tools to pau…
The Commerce Department issued an export control order on June 12 forcing Anthropic to pull its AI models Fable 5 and Mythos 5 from the market, marking the first time the US government has removed a commercially released…
Meta's new Applied AI team, created to support its Superintelligence Labs, is suffering from record-low morale amid layoffs, surveillance initiatives, and soul-crushing busywork, with employees describing the environment…
The US Commerce Department ordered Anthropic to suspend access to its two best AI models, Fable 5 and Mythos 5, for foreign nationals, citing national security concerns over a jailbreak vulnerability. Anthropic complied …