X to introduce AI-generated Community Notes in pilot program

X (formerly Twitter) has announced it will allow artificial intelligence to author Community Notes, marking a significant expansion of its fact-checking system.

Key Details:

• Pilot Program Launch: Begins today with developer tools for creating AI Note Writers
• Initial Limitations: AI-written notes will only appear as replies in test mode and require approval
• Public Rollout: First AI-authored notes expected to appear later this month
• Transparency: All AI-generated notes will be clearly labeled

How It Works:

1. AI Note Writers will be assessed by an open-source, automated note evaluator that checks for:
   • Relevance to the topic
   • Potential harassment or abuse
2. The evaluator bases decisions on historical input from human Community Notes contributors
3. Initial AI notes will only respond to posts that specifically request Community Notes

Company's Vision:

X positions this move as creating "a powerful feedback loop" where:

• AI accelerates the speed and scale of fact-checking
• Community ratings help improve AI accuracy and reduce bias

The announcement emphasizes "humans still in charge", though the primary human oversight appears to come from community ratings rather than pre-publication editorial review.

Potential Challenges:

• Effectiveness remains unproven (Community Notes has faced criticism for failing its core mission)
• Questions about how the automated evaluator determines "bias"
• Long-term plans suggest expanding AI's role in the fact-checking system

Developers can access the Community Notes API guide for technical details about creating AI Note Writers.