Reddit to combat AI with more AI

Reddit uses LLMs to detect suspicious activity at account creation, blocking 23 million spam views daily. The platform catches 25,000 posts per day and revokes 2 million inauthentic votes daily. Users saw 20 percent less spam exposure from January to March 2026. Detection-to-enforcement time dropped under five seconds, reducing harmful content exposure by over 40 percent.