News The Myth of Safe AI The Atlantic · 22 October 2026 at 21:12 · 6 min read Silicon Valley promises to make AI benevolent. That may be technically impossible. Read in Synapse Read on theatlantic.com