GitHub Explores Pull Request Limits to Curb AI-Generated Code Flood
GitHub is considering implementing tighter controls on pull requests (PRs) to deal with the explosive growth of AI-generated contributions, according to reports from InfoWorld and AlphaSignal.
The current numbers are staggering:
- GitHub processes 275 million commits per week
- On track for 14 billion commits in 2026 (vs 1 billion in 2025)
- AI agents now participate in 14.9% of PRs, up from just 1.1% in February 2024
- Growth driven by tools like Cursor, Claude Code, GitHub Copilot, and Codex
The problem: much of this AI-generated code is low quality. Open-source maintainers (who are already chronically under-resourced - 94% of widely-used projects have fewer than 10 developers) are receiving avalanches of automated PRs that require manual review.
Solutions GitHub is evaluating include: per-contributor PR limits, AI-based filters for detecting low-quality patches, and tools for maintainers to set stricter contribution policies.
The shift reflects a growing dilemma in the open-source ecosystem: AI agents can accelerate development, but they can also overwhelm the human review infrastructure that maintains code quality.