This project explores how AI can help write Community Notes—short, cited fact-checks that provide context on social media posts. The challenge is straightforward: misinformation spreads at the speed of social media, so fact-checking must keep pace. By combining AI's scalability with the transparency of Community Notes, we're testing whether it's possible to deliver better information exactly where misleading claims are circulating.

Our system monitors posts where Community Notes have been requested, automatically drafts potential notes, and makes them available for review. The system draws from reliable sources, maintains a neutral tone, and covers topics ranging from politics and technology to health, science, and breaking news. Our goal is to refine this process so that generated notes are not only fast, but consistently helpful and trustworthy.

X Community Notes Program

This work operates within X's Community Notes API program, which X.com recently opened to enable AI-generated Community Notes. We work within X's established framework, follow their guidelines, and submit notes through their API alongside other program participants.

Early Beta Status

This is an early beta project, and we expect to make mistakes—potentially many of them. Fact-checking is inherently challenging: sources conflict, language carries nuance, and context matters deeply. Automating this process with AI compounds these difficulties. Some of our notes will be incomplete, confusing, or incorrect.

Despite these challenges, the potential impact makes the effort worthwhile. If we can develop this approach successfully, even incrementally, it could provide a scalable method for improving information quality online. We're committed to learning from our errors and continuously improving the system.