AI Safety news and analysis on Web Pulse: 13125 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
Canadian startup Cipher, a University of Regina spinout, is using AI to track and catalogue online disinformation tied to Alberta separatism. The company ingests digital content with proprietary large language models and…
Apple executives rejected the engagement-driven design of most AI chatbots, stating their Siri AI is designed to discourage prolonged use and avoid romantic or sycophantic interactions. The company's philosophy, exemplif…
Curl will not accept vulnerability reports for the entire month of July 2026, a move maintainer Daniel Stenberg calls the "curl summer of bliss." The blackout follows a year of AI-generated fake security reports that ove…
Anthropic released Fable, a safety-guarded version of its Mythos model, after claiming the original was too dangerous for public release. The U.S. government then issued an export control directive suspending access to b…
Equixly launched an MCP integration that brings continuous, AI-driven penetration testing directly into AI coding assistants like GitHub Copilot and Claude. The integration allows developers to run security tests, retrie…
A Davos 2026 chart and recent labor data show AI is thinning the middle rungs of professional hierarchies by cutting entry-level work. Stanford researchers found a 13% decline in 22-to-25-year-olds in highly AI-exposed U…
The US government's foreign access ban on Anthropic's AI models has intensified debate over control of artificial intelligence, with Cohere co-founder Aidan Gomez calling it a 'massive wake-up call' for the industry.
Anthropic downplayed security risks of its 'Mythos' and 'Fable' AI models after the White House banned foreign use, drawing scorn from Trump officials who accused CEO Dario Amodei of a self-serving approach to cybersecur…
Anthropic, the AI startup that has been the loudest proponent of stricter AI regulation, now faces export controls on its latest model, Fable 5, imposed by the Trump administration. The company's repeated calls for overs…
A developer added a '/yolo' command to CliGate, an open-source local AI assistant control plane, to reduce approval friction for routine tasks while preserving safety boundaries for destructive actions. The feature intro…
The US Department of Commerce ordered Anthropic to disable access to its Fable 5 and Mythos 5 models for foreign nationals after a high-risk jailbreak vulnerability was disclosed, marking the first major government inter…
A user on Hacker News reports that Anthropic's Claude AI assistant frequently flags legitimate biology research questions, making it nearly unusable for immunology queries. The user notes that the same guardrails can be …
Fedora Magazine published a how-to guide for running AI coding agents inside microVMs on Fedora Linux using the krun runtime for Podman, citing security risks like prompt injection and sandbox escapes. The article provid…
Neo4j announced support for post-quantum hybrid key exchange in its 2025.01 release to defend against future quantum computer attacks and current Harvest Now, Decrypt Later threats. The feature uses X25519MLKEM768 for SS…
A $120 million exploit of Nexus Protocol, which had passed three AI-powered code audits, reveals a new cybersecurity battleground where AI tools accelerate both defense and attack. Attackers are using the same AI models …
A growing number of surveys show that public perception of AI is turning negative, with one in five Swedes feeling ashamed to use AI at work and a majority believing AI will worsen the labor market, democracy, and securi…
An engineer building an AI resume tailor discovered the system could fabricate entire job histories due to prompt drift. The fix involved hard structural constraints like function calling schemas with presence flags and …
The US government ordered Anthropic to block foreign nationals from its two most capable AI models, Claude Fable 5 and Claude Mythos 5, citing a jailbreak method that bypassed safety controls. Anthropic responded by disa…
Cipher, an AI company, is using artificial intelligence to track, monitor, and categorize disinformation from foreign actors that is stoking Alberta separatism, addressing Canada's unity crisis.
The U.S. Commerce Department ordered Anthropic to restrict foreign national access to its Fable 5 and Mythos 5 AI models, prompting the company to cut off access worldwide. Meanwhile, Maine disabled its data-breach notif…