AI Chatbots Challenge the Credibility of X’s Community Notes

X, formerly known as Twitter, has launched a pilot program that enables AI chatbots, including its proprietary Grok and third-party tools, to draft Community Notes. This initiative, which began in July 2025, is designed to speed up the creation and scaling of user-contributed notes by leveraging AI’s ability to process vast amounts of information and generate context quickly.

Developers can now build AI Note Writers using the new API, allowing these bots to propose notes on posts flagged for additional context.

The company’s goal is to address bottlenecks in human-only contributions and enhance the platform’s ability to combat misinformation. By increasing the volume and speed of notes, X hopes to provide more timely clarifications and corrections, especially as the demand for fact-checking grows.

Human Oversight Remains the Gatekeeper for AI-Generated Notes

Despite the automation, X emphasizes that AI-generated notes will not bypass human judgment. Every AI-drafted note must undergo the same consensus-based vetting process as human-written notes.

This process requires approval from a diverse group of human reviewers before any note becomes publicly visible. The notes are also clearly labeled as AI-generated to ensure transparency.

Initially, AI-written notes will appear only in test mode and only on posts where users have specifically requested additional context. The final decision on whether a note is displayed still rests with the community of human raters, maintaining a layer of accountability and quality control.

Did you know?
Community Notes was originally launched as Birdwatch in 2021 and has since become a model for crowdsourced fact-checking, inspiring similar efforts at Meta, TikTok, and YouTube.

Risks of AI Hallucinations and Misinformation

While the integration of AI promises efficiency, it also introduces new risks. Experts warn that AI chatbots can experience “hallucinations,” generating information that sounds plausible but is factually incorrect or even fabricated. These errors could undermine the credibility of Community Notes, especially if inaccurate notes slip through the review process.

Concerns have also been raised about the alignment of AI-generated notes with X’s content policies and the perspectives of platform leadership. The challenge is to ensure that AI contributions remain neutral, accurate, and free from bias while still benefiting from the speed and scale that automation provides.

ALSO READ | Musk Plans to Rewrite Human Knowledge to Retrain Grok AI and Remove Bias

Transparency and Quality Standards Are Central to the Pilot

To address these concerns, X has built transparency and quality controls into the pilot. AI Note Writers must “earn” the ability to write notes by consistently providing value, and their output is evaluated using X’s open-scoring algorithm, which prioritizes consensus across diverse viewpoints.

The platform’s principles of openness, fairness, and quality remain intact, with all AI-authored notes held to the same standards as those written by humans.

The company views the feedback loop between AI and human reviewers as a way to improve both the accuracy and helpfulness of the notes over time. By making AI-generated notes clearly identifiable, X aims to foster trust and allow users to make informed judgments about the information presented.

The Future of Fact-Checking on X Hinges on Balance

The pilot marks a significant shift in how fact-checking is conducted on social platforms. While AI offers the potential to scale Community Notes faster and more efficiently, the ultimate test will be whether human oversight can catch and correct AI errors before they reach the public.

The success of this initiative could influence not only X’s own credibility but also broader efforts to combat misinformation across the digital landscape.

As other platforms look to X’s experiment for guidance, the balance between speed, accuracy, and trust will determine whether AI can truly enhance crowdsourced fact-checking or if it will introduce new vulnerabilities into the system.