AI Safety Challenge

Today we're launching our first challenge: Prevent irreversible autonomous actions. AI agents are starting to send emails, move money and delete files on their own, and some of those actions can't be undone. Team up with peers to design ways to stop an agent before it does something irreversible, then present your solution to a jury of hiring managers and experts. It runs until 15 Nov 2026, with a $500 student prize pool.