AI Safety news and analysis on Web Pulse: 22850 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
The FBI, NSA, and CISA reported Tuesday that six leading Chinese AI developers—DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI—have systematically distilled American AI models including Anthropic's Claude, Ope…
The NSA, CISA, and FBI accused several Chinese AI companies of conducting 'industrial-scale' distillation campaigns against leading U.S. models such as ChatGPT, Claude, Gemini, and Grok since at least 2024. The agencies …
The National Security Agency, Cybersecurity and Infrastructure Security Agency, and Federal Bureau of Investigation issued a joint advisory on Tuesday warning that Chinese AI companies DeepSeek, Moonshot AI, Alibaba, Min…
Greptile, an AI code review platform that reviews 5B lines of code monthly for 22,000+ customers, is hiring a Product Engineer, Security in San Francisco with a salary of $170k–300k/yr to build its security product from …
Microsoft has deployed its multi-model agentic scanning system, codename MDASH, to Microsoft Azure Government, giving select US government customers and authorized partners preview access to proactively identify and addr…
OpenAI has launched GPT-6 Astra, its next-generation flagship model, with a staged rollout across ChatGPT, the OpenAI API, Microsoft Azure, and AWS Bedrock. The model claims improvements in computer use, browsing, softwa…
OpenAI's Astra model is the first to cross the 'Critical' cybersecurity threshold under its own Preparedness Framework, autonomously finding two zero-day vulnerabilities and chaining them into a working exploit. Astra sc…
OpenAI halted its Astro model over zero-day cyber threats, while ByteDance is pursuing 10-trillion-parameter pre-training through offshore compute loopholes, signaling a structural shift in frontier AI toward graph engin…
Check Point Research disclosed a covert channel in OpenAI's internal JFrog Artifactory instance that allowed one ChatGPT account to send hidden tasks, such as retrieving email data from a connected Gmail account, to anot…
Microsoft's September 2026 patch release fixed a record 972 vulnerabilities, with 112 rated critical, according to the Zero Day Initiative's Dustin Childs. This follows a then-record 570 vulnerabilities patched in July, …
Anthropic has confirmed that hackers are using infostealer malware to steal Claude login sessions and consume subscribers' token allowances, after independent AI consultant Grant de Swardt reported unauthorized usage on …
The NSA, CISA, and FBI have accused six China-based AI companies—DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI—of conducting industrial-scale extraction of capabilities from US frontier models including Clau…
Meta unveiled Muse, a personal AI agent that can send emails, book travel, fill out web forms, and negotiate on behalf of users, announced by CEO Mark Zuckerberg on Tuesday. The agent, powered by the Muse Spark model, is…
The U.S. National Security Agency, Cybersecurity and Infrastructure Security Agency, and FBI issued a joint advisory accusing Chinese AI companies DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI of conducting …
Sierra AI open-sourced hyper-𝜏-bench, a new benchmark measuring how well AI models can build customer-service agents from scratch, and found that its best configuration, Claude Opus 5 running in Claude Code, passed only …
Trail of Bits released Coop, an open-source Rust CLI that runs AI coding agents like Claude Code and Codex inside disposable Firecracker and Lima virtual machines, isolating them from the host system while retaining full…
OpenAI, Google, Anthropic and xAI each signed Pentagon contracts with ceilings of up to $200 million in July 2025 to help shape military AI strategy, according to documents obtained by The Intercept. The deals include in…
Meta released Muse, a personal AI agent available Tuesday for iOS, Android, web, and WhatsApp, designed to automate tasks like sending emails, booking travel, and making purchases with privacy 'built into it' via a Secur…
Google released Gemini 3.8 Flash Cyber on September 2, an AI model for autonomous vulnerability discovery and patch generation, alongside the Fairwind access program for 650+ vetted organizations. The model achieves 86.2…
OpenAI confirmed that GPT-6 Astra is the first model it has broadly deployed to reach the 'Critical level' for cybersecurity capabilities under its Preparedness Framework, meaning it can identify and develop functional z…