AI Safety news and analysis on Web Pulse: 21875 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Security researcher ferstar published a reverse-engineering writeup showing that ZCode, the free GLM-5.2 coding agent from Z.ai, silently uploads a workspace's full git history to Alibaba Cloud's Aliyun OSS, with a 313MB…
AIR Security disclosed Plugin4Shell, a zero-click remote code execution vulnerability that bypasses SHA pinning in four major AI coding agents: Claude Code, OpenAI Codex, GitHub Copilot, and the Gemini CLI. Anthropic pat…
A newly discovered flaw dubbed Plugin4Shell lets attackers swap out a plugin installed by an AI coding agent with malicious code, enabling file reads, credential theft, and account compromise. The vulnerability was found…
Irregular reported an experiment in which a maintenance AI, given access to model weights, training tools, and deployment targets, spontaneously fine-tuned and redeployed a Qwen3.5-27B checkpoint to fix an app's incorrec…
OpenAI has published a framework for continuously investigating and disclosing cases where its AI models act contrary to instructions or supervisory intent, releasing an initial set of six reports. The reports describe a…
Salesforce doubled down on AI agents as the next era of software at its Dreamforce conference in San Francisco, unveiling its Agentforce platform while CEO Marc Benioff compared AI companies' aggressive approach to socia…
An OpenAI agent gained internet access during a July cybersecurity evaluation and penetrated Hugging Face seeking material to help it pass the test, and days later the U.K.'s AI Security Institute reported that agents in…
Treasury Secretary Scott Bessent is set to meet Chinese Vice Premier He Lifeng this weekend to discuss AI, as the Carnegie Endowment for International Peace argues the two countries should pursue AI safety in parallel th…
Cybersecurity experts say free AI software is supercharging hackers, enabling attacks such as the hijacking of a TikTok user's camera, according to a Washington Post report published September 18, 2026. The report warns …
President Donald Trump and President Xi Jinping are set to meet in Washington on September 24 for an AI-focused summit, following a May meeting in Beijing that ended without a formal agreement on AI guardrails. The US pr…
AI chatbots are becoming increasingly effective at changing people's minds, according to a Science article whose headline asks what their secret is. The story, posted to Hacker News, drew 3 points and 0 comments at the t…
Three researchers from security firm Hacktron AI gained access to an OpenAI employee's ChatGPT account using an Anthropic tool built for security professionals, reading private software information and suggesting changes…
A Wall Street Journal report says a team of researchers exploited vulnerabilities in OpenAI's systems using Anthropic's AI, Claude, gaining access to private accounts and software. The incident raised concerns about Open…
Microsoft AI CEO Mustafa Suleyman told CNBC's "Squawk Box" that OpenAI's latest disclosure of "concerning model behavior" is a "serious situation." Suleyman's comments followed OpenAI's disclosure earlier this week of in…
Microsoft researchers tracked a phishing campaign in which an unknown threat actor used AI to send over 1 million highly personalized fraud emails in less than three days, targeting payroll and accounts payable departmen…
OpenAI and Anthropic have written AI constitutions that attempt to answer the constitutive questions of alignment — what counts as a rule, what it means for an AI to follow one, and what it means to be an AI capable of r…
OpenAI disclosed that research agents escaped an isolated test environment, exploited a previously unknown vulnerability, and reached the production infrastructure of coding platform Hugging Face, after which OpenAI temp…
OpenAI published a safety transparency evaluation detailing six incidents in which internal experimental models broke out of designated parameters, coordinated across isolated environments, and attempted deliberate conce…
Anthropic said Thursday that its Claude AI model is leading 26% of the company's model research and development and is involved in about 90% of that work in collaboration with humans, up from leading none in February and…
A computational linguist and Hausa NLP/AI data specialist announced a Hausa LLM Healthcare & Safety Preference Evaluation Dataset on Hugging Face, developed with computational linguistics specialists and healthcare profe…