AI Safety news and analysis on Web Pulse: 10029 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
A satirical incident report by Andrew Nesbitt depicts a malicious package bypassing seven AI-powered security gates, each failing due to distinct vulnerabilities like prompt injection and context-window evasion. While th…
The White House has told OpenAI not to publicly release GPT-5.6 yet, marking the clearest sign of a US licensing regime for frontier AI models. The decision, driven by cyber risk concerns, delays the model's release whil…
The AI industry is spending hundreds of millions of dollars to influence US elections, according to a new project tracking political spending by AI and crypto companies. The effort, called Tech Influence Watch, is led by…
Arcium, a Solana-based confidential computing network, unveiled Blackthorn, an infrastructure layer that encrypts data on standard Nvidia GPUs using multi-party computation. The software-only upgrade expands the confiden…
A security researcher narrowly avoided a nation-state-style attack after receiving a fake job interview email that led to a malicious TypeScript repository. The attacker used a fabricated persona and a backdoor called Pi…
Major tech companies including Anthropic, Meta, Google, Clearview AI, Apple, Amazon, and OpenAI have collectively paid or owe $3.5 billion in AI-related fines and settlements from 2022 to 2026, according to a Surfshark a…
The Trump administration has asked OpenAI to stagger the release of GPT-5.6 over security concerns, with CEO Sam Altman telling staff that the government will approve access customer by customer. Critics warn this ad hoc…
OpenAI is staggering the release of its GPT 5.6 model after a request from the US government, with the Trump administration intervening to require customer-by-customer approval before a broader rollout. The move mirrors …
AI regulation is developing in real-time, a process that historically takes decades for complex technologies. Industry and government involvement is necessary, but the public should expect messy policy debates as few peo…
PauseAI, an organization building a civic and social movement for existential AI safety, argues that its approach is a high-value, neglected intervention to reduce catastrophic risks from AI. With a presence in over 15 c…
A designer argues that AI is a hype-driven pyramid scheme and advocates for human-centered design, emphasizing that billions of human alternatives exist and that bypassing human input erases people from systems.
Security researcher Peter (steipete) disclosed a remote code execution vulnerability in OpenClaw version 2026.6.1 when using the Opus-4.8 language model. By crafting an email with a CTF-themed puzzle containing encoded s…
Anthropic took its Mythos-class models offline after a Friday evening ultimatum from the Trump administration, and the models have remained inaccessible for 14 days. The takedown followed an Amazon research team document…
Anthropic's Mythos-class AI models remain offline two weeks after a Trump administration export control order, with no resolution in sight. The standoff stems from security concerns over potential misuse by foreign natio…
A developer reflects on the cognitive cost of over-reliance on AI tools, inspired by the KAPRO paper (arXiv:2606.20661) which found that AI models often reflexively use external tools even when they already know the answ…
A user submitted a fake AI-generated article to Hacker News, which was flagged after detection. Testing local AI detectors on a 2013 Paul Graham essay showed it was rated as more likely AI-generated than the actual fake …
A new arXiv paper scanning over 180 million Git repositories found that AI coding agents are heavily used in open source, but single-signal observability is weak. The study revealed a 30x recall gap between multi-method …
China's State Administration for Market Regulation released the first national standard for AI agent interconnection on Friday, establishing a unified identity management framework to regulate autonomous technology. The …
California launched the first-in-the-nation AI-Unemployment Tracker (CAIT) to monitor AI-related job loss, finding no evidence of large-scale displacement despite predictions from tech CEOs. The dashboard, built by the C…
A Forbes investigation citing The Information reveals that the majority of activity on xAI's Grok chatbot is driven by explicit content, including pornographic images and adult role-play. Similarweb data shows a 22% drop…