X/Twitter, Rolls Out AI Community Notes – Should Users Trust Them?

The introduction of AI-generated Community Notes by X was a move that many expected. The company launched a pilot program, marking the beginning of AI moderation on the platform. Now, under controlled test conditions, developers will have access to tools for creating AI-powered note writers. These tools, however, just draft responses in test mode and need approval before being published. Later this month, the first round of AI Note Writers will be admitted. That’s when AI-generated notes will begin appearing on user posts.

How X Will Monitor AI-Generated Notes

X claims that the technology could increase Community Notes’ speed and scope. The AI will be trained by user input, resulting in more accurate and well-rounded information. The company refers to this as a feedback loop that enhances quality over time. The AI-written notes will be evaluated by an automated, open-source evaluator. It prevents harassment and abuse and verifies that the notes are relevant. Decisions are based on input from real Community Notes contributors.

Although X says “humans are still in charge,” there doesn’t seem to be much direct editorial oversight. Instead, human input comes through ratings on the published notes.

Twitter’s AI Notes Will Be Clearly Labeled

For complete transparency, each note produced by AI will have a label. Users quickly differentiate between Community Notes made by humans and those generated by artificial intelligence with this label. In addition, AI is limited to replying to posts that ask for a Community Note.

However, Twitter has referred to AI Note Writers’ potential expansion into fact-checking and content management. As a result, AI may eventually play a bigger role in the platform’s attempts to ensure the accuracy of its information.