Related ToolsCursorClaude CodeCodyAider

GitHub Will Use Copilot User Data for AI Training Starting April 24

Microsoft Copilot
Image: Microsoft

GitHub just reversed its position on AI training data. Starting April 24, the company will collect and use interaction data from Copilot Free, Pro, and Pro+ users to train its AI models.

The data collected goes well beyond what most users probably expect. It includes accepted and modified code suggestions, the code context around your cursor, comments and documentation, file names, repository structure, chat interactions, and even thumbs up/down feedback. Private repository content is also fair game while you're actively using Copilot.

Who's Affected and Who's Not

Business and Enterprise customers are carved out entirely - their contracts protect them. Students and teachers on GitHub Education are also exempt. That leaves individual developers on Free, Pro, and Pro+ plans as the training data source.

To opt out, go to /settings/copilot/features and uncheck "Allow GitHub to use my data for AI model training" under Privacy settings. If you don't act before April 24, you're in by default.

The Reaction Is Exactly What You'd Expect

GitHub opened a community discussion about the change. Out of 39 posts, exactly one person endorsed the new policy - Martin Woodward, GitHub's VP of Developer Relations. The emoji reactions tell the story more bluntly: 59 thumbs-down versus 3 rocket ships.

GitHub CPO Mario Rodriguez pitched it as a way to "help our models better understand development workflows, deliver more accurate and secure code pattern suggestions, and improve their ability to catch potential bugs before production." Standard corporate framing for "we need more training data."

This follows a pattern. GitHub Copilot's underlying model was originally trained on publicly available GitHub code without explicit user consent. That already sparked legal challenges. Now the company is formalizing what amounts to a broader data collection policy, just with an opt-out toggle.

For developers who care about code privacy, the calculus is simple: either opt out before April 24, switch to a Business/Enterprise plan, or start evaluating alternatives. The opt-out exists, but making it opt-in by default is a deliberate choice - GitHub knows most people won't change their settings.