AI Safety & Policy
AI safety, alignment, and governance
Articles
A New Chatbot Wants to Unlock the Secrets in Tattered Ancient⭐9
Researchers built an ancient Greek language model that fills gaps in damaged papyrus fragments, achieving 62% accuracy to help historians reconstruct lost texts...
AI, Tariffs, and Rare Minerals: What to Expect From the Trump–Xi⭐9
What to expect from the Trump–Xi summit: AI chip export controls, tariffs, and rare earth minerals as bargaining chips in US-China talks.
US and China Discuss Mutual Alert System for AI National⭐8
The US and China discuss a mutual alert system for AI incidents threatening national security, signaling rare cooperation on AI risk management.
Donald Trump vs. MAGA: The Growing Rift Over Data Centers⭐9
As Trump pushes AI data centers, a growing rift with his MAGA base emerges over energy costs, water use, and local backlash in red and battleground states.
Meta's Muse Is Better at Surveilling Than Helping Me⭐8
Meta's Muse app opts users into AI training data collection by default and nudges them to share bank, email, and passport details.
Forget the AI Slowdown—the Vulnerability Explosion Is Already⭐8
AI labs debate slowing development, but existing chatbots already fuel a vulnerability explosion, overwhelming security teams. Here's what it means for 2026.
Mathematicians Feel Threatened by AI—But They Can't Quit It⭐9
AI is transforming mathematics, proving theorems and generating conjectures, leaving mathematicians anxious about their future yet unable to quit tools they rel...
Join the WIRED World Fair in Miami on November 4⭐8
Join WIRED World Fair in Miami on November 4, 2026, for live demos, immersive tech, and talks with industry leaders shaping the future.
Here's How an AI Slowdown Could Actually Be Enforced⭐8
As AI capabilities accelerate, enforcing a coordinated slowdown faces steep hurdles—from verifying secret training runs to international coordination and corpor...
If the AI Industry Followed Its Own Research, It Might Have⭐10
Anthropic's CEO says AI safety requires understanding how models "think." Interpretability research suggests they don't—and that by its own logic, the industry ...