Reddit Deploys AI Tools to Combat AI-Generated Marketing Spam Targeting Chatbots

Reddit has deployed advanced large language models to combat a rise in AI-generated marketing spam, known as generative engine optimization, which targets the platform to manipulate chatbot search results. From January to March 2026, Reddit's automated systems reduced spam exposure by 20%, blocking 23 million spam views and revoking nearly 2 million fake votes daily. Additionally, automated enforcement against English-language hate and violence now takes under five seconds, reducing exposure by over 40%.

AI-powered spam detection systems

• ▪Marketers use generative engine optimization to plant fake opinions on Reddit, aiming to influence recommendations from artificial intelligence chatbots
• ▪Reddit deployed large language models to detect subtle, coordinated patterns of fake behavior and artificial hype on the platform
• ▪Reddit analyzes signals at the time of account creation to identify and block suspicious actors before they can post
• ▪Reddit requires automated accounts flagged as suspicious to verify their humanity

Automated content moderation results

• ▪Reddit reduced user exposure to spam by approximately 20% from January to March 2026 compared to the prior three months
• ▪Reddit's automated systems block 23 million spam views per day before they reach users
• ▪Reddit's automated systems catch approximately 25,000 net new spammy posts and comments daily
• ▪Reddit's automated systems revoke nearly 2 million inauthentic votes per day

Harmful content enforcement improvements

• ▪Reddit expanded its automated systems to enforce policies against hate and violence in all English text content
• ▪Reddit reduced user exposure to potentially harmful content by more than 40% through faster automated enforcement
• ▪Reddit increased enforcement actions on hate and violent content by more than 200%
• ▪Reddit reduced the average time between detection and enforcement on harmful content containing hate or violence to under five seconds

Reddit's layered moderation approach

• ▪Volunteer community moderators were responsible for more than half of post and comment removals on Reddit between July and December 2025
• ▪Reddit utilizes a layered moderation approach consisting of platform-level administrators, volunteer community moderators, and user upvoting and downvoting