AI Safety news and analysis on Web Pulse: 10914 curated articles tracking the latest AI Safety developments, tools, and research, updated continuously from vetted sources.
A new analysis suggests that AI systems may be making leaders more dangerous by enabling overconfidence and reducing accountability, contrary to the expectation that AI would improve decision-making.
The United States government is pressuring Meta to submit its artificial intelligence systems for federal review amid growing national security concerns, according to sources familiar with the discussions. The push refle…
A paper by Saha et al. (arXiv 2506.07001) demonstrates that detector-guided paraphrasing using RoBERTa as a reward reduces the true positive rate of AI-generated text detectors by 87.88% across Binoculars, Fast-DetectGPT…
New York Assemblyman Micah Lasher won the Democratic primary in Manhattan's 12th Congressional District to succeed retiring Rep. Jerrold Nadler, NBC News projects. The race drew heavy outside spending from AI-aligned gro…
A new collection by Andrew Maynard, director of Arizona State University's Future of Being Human initiative, examines historical technology-driven moral panics from Ancient Greece to modern AI, arguing that these reactio…
US Treasury Secretary Scott Bessent said the biggest risk in artificial intelligence is China getting ahead of the United States, outweighing concerns over safety or job losses. He warned of retaliation against any attem…
Legion LegalTech Corp sued the US government on June 23, 2026, challenging a Bureau of Industry and Security order that forced Anthropic to block foreign access to its Fable 5 and Mythos 5 AI models, disrupting the compa…
Researchers stress-tested Natural Language Autoencoders (NLAs) by optimizing activation vectors to flip AV explanations while preserving model behavior, achieving an 81.4% flip rate with 99.6% label preservation. Further…
The White House met with children's safety groups to discuss a package linking online safety bills to a federal AI framework, including the Kids Online Safety Act and the App Store Accountability Act. House leaders annou…
AI-focused super PACs have raised over $100 million and spent $49 million on the 2026 midterm elections to influence AI regulation. The Manhattan Democratic primary in New York's 12th district alone has attracted more th…
The Trump administration is pressing Meta to voluntarily submit its AI models for government security reviews, making it the only major US AI company without such an agreement. The push follows an executive order establi…
Palantir CEO Alex Karp co-wrote a manifesto titled 'The Technological Republic,' which journalist James Corbett reviewed as unhinged. Palantir, founded by Peter Thiel and backed by the CIA, develops AI-powered surveillan…
Cato Networks joined OpenAI's Daybreak Cyber Partner Program on June 22, 2026, integrating GPT-5.5 and Codex security tools into its SASE platform to reduce vulnerability response times. The partnership aims to compress …
AI systems inevitably make unpredictable mistakes, requiring verification of every output. This verification often costs as much as doing the work manually, leading companies to externalize that cost onto workers, who be…
An 18-year-old suspect, Bradley Scott Sayer, fatally shot two men at a Chico, California library on Monday, officials said. The suspect wore a Columbine-inspired shirt and was arrested within minutes of the first 911 cal…
The next wave of AI systems must focus on stewardship after task completion, not just clever responses. The real challenge lies in making AI loops inspectable, recoverable, and governable for ordinary humans. The future …
Anthropic's Mythos AI model identified vulnerabilities in highly sensitive U.S. government computer systems during a testing exercise, with Senator Mark Warner claiming it broke into almost all classified systems within …
Legal-tech startup Legion sued the U.S. government over a Commerce Department directive that prompted Anthropic to disable access to its top-tier AI models Fable 5 and Mythos 5, causing immediate harm to Legion's Canada-…
Meta paused its employee monitoring program after internal security breaches allowed unauthorized access to sensitive data, including keystrokes, screen content, and private conversations. The program, designed to train …
A developer's platform for structured compliance-training programmes repeatedly failed internal quality checks on an ISO 42001 Lead Auditor course due to subtle drift into other standards. Investigation revealed 22 disti…