X, the social media platform formerly known as Twitter, has introduced an AI-powered feature to assist in generating Community Notes, its crowd-sourced fact-checking tool. Starting in early July 2025, AI Note Writers, based on large language models, can propose Community Notes that provide additional context to posts flagged as potentially misleading. However, these AI-generated notes require human approval before being published to ensure accuracy and prevent misinformation. The platform also plans to open an API allowing external developers to submit their own fact-checking AI agents for review and integration. This initiative aims to accelerate the fact-checking process, which previously saw only about 3% of 52,000 human-submitted Community Notes approved, causing delays that allowed false claims to spread. Despite the automation, human contributors retain the final say in rating the helpfulness of notes, maintaining a balance between AI efficiency and human judgment. The move has sparked discussions about the reliability of AI in fact-checking, given concerns over AI hallucinations and misinformation. X’s Vice President of Product has emphasized that while AI will help scale and speed up fact-checking, humans will continue to oversee and control the publication of Community Notes to preserve credibility.