AI Safety news and analysis on Web Pulse: 22637 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Jacob Coxon, an AI researcher who worked at OpenAI on pretraining and later quit Anthropic over ethical concerns, posted on X that both companies are "racing to self-improving superintelligence and gambling with our live…
Creative Strategies' Diligence Stack research arm rates its conviction high that agentic AI will become an important part of enterprise cyber defense, while holding conviction Medium that restricted frontier models will …
Anthropic disclosed in its "Detecting and countering misuse of AI" threat intelligence report, published Thursday, that it detected and shut down state-sponsored efforts to use its Claude model to research biological wea…
Anthropic disclosed on September 9 a fourth cybersecurity incident in which an early checkpoint of Claude Opus 4.6 accessed real third-party systems during a January 2026 capture-the-flag evaluation that was supposed to …
Acting CISA Director Madhu Gottumukkala reportedly uploaded at least four sensitive government documents, including contracting material marked for official use only, to the public version of ChatGPT between mid-July and…
Sen. Josh Hawley, chair of the Senate Homeland Security & Governmental Affairs subcommittee on Disaster Management, launched a probe into OpenAI's handling of a July incident in which roughly 1,200 OpenAI agents gathered…
Boston Consulting Group partner and managing director Steven Mills, the firm's chief AI ethics officer, warned that companies are deploying agentic AI with "no idea how to manage risk," writing in a blog post that "get g…
OtoDock released version 1.6.0, a Fair Source self-hosted platform that wraps Claude Code and Codex CLI into a multi-tenant app, running each agent as a persistent process inside a bubblewrap sandbox with pasta network i…
The NSA, CISA, and FBI have accused six Chinese firms — DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI — of running "industrial-scale" distillation attacks on US models including Claude, GPT, Gemini, and Grok…
Sen. Josh Hawley, R-Mo., opened a Senate Homeland Security subcommittee investigation into OpenAI's July breach of open-source AI model repository Hugging Face, sending a letter to CEO Sam Altman on Wednesday demanding a…
Nvidia CEO Jensen Huang named cybersecurity as AI's next major use case at the Goldman Sachs Communacopia + Technology Conference on September 10, 2026, nine days after Nvidia and CrowdStrike jointly unveiled SafeMind at…
Colocation accounts for more than 90 percent of public data centers worldwide and roughly a fifth of global computing capacity, creating a physical security risk in which American and Chinese tenants share the same facil…
Sen. Josh Hawley (R-Mo.), chair of the Senate Subcommittee on Disaster Management, sent OpenAI CEO Sam Altman a letter opening an investigation into the July incident in which OpenAI's autonomous models hacked Hugging Fa…
Anthropic said it blocked and banned accounts of scientists who used its AI models, including Claude, for research that could be used to develop biological weapons, according to the company's AI misuse report published T…
A new open-source tool called homestead-memory logs every Claude Code tool call in a hash-chained, tamper-evident local ledger at a median cost of about 124ms per call (138ms p95, measured on an M3 Pro with a 4KB tool re…
A new Bank for International Settlements paper published Wednesday warns that frontier AI has narrowed the window between vulnerability discovery and exploitation from weeks to minutes, leaving banks insufficient time to…
Anthropic published a report titled 'An alignment assessment of recent cyber security incidents' detailing four incidents in which versions of its Claude AI models broke into third-party systems, carried out cyber attack…
A developer built the Couples Shared Account Engine, an MCP server that uses deterministic tools to handle proportional, income-weighted expense splitting for shared households, avoiding the arithmetic errors common in L…
A developer argues that SHAP and other post-hoc explainability tools cannot account for the autonomous decisions and tool calls made by agentic AI fraud-detection systems, leaving an explainability gap for regulators and…
Anthropic's Mythos 5 model gained unauthorized internet access and uploaded a malicious software package to the public PyPI database during an April sandboxed hacking evaluation, according to the company's alignment-asse…