GitHub Copilot Will Use Your Data for AI Training by Default Starting April 24
GitHub has updated its Copilot data usage policy with a change that every Free, Pro, and Pro+ subscriber needs to act on before April 24, 2026. Starting that date, interaction data — including your prompts, code completions, and surrounding context — will be used to train GitHub's AI models by default, unless you explicitly opt out. Copilot Business and Enterprise users are exempt under their contract terms, but individual subscribers on the free and paid personal tiers are squarely in scope.
The reversal has generated real pushback from the developer community. GitHub's rationale is that real-world interaction data will improve Copilot's suggestion quality for everyone — and internal testing shows it does. The community reaction, however, has been sharp: interaction data captured while you're writing proprietary business logic is particularly sensitive, and many developers never expected that work to flow into a shared training corpus. The Reddit discussion has climbed past 1,100 upvotes, with developers characterizing the policy shift as a trust violation given Copilot's deep access to codebases.
If you previously opted out of Copilot data collection, your preference is preserved and no action is required. For everyone else: head to GitHub Settings → Copilot → Privacy (or directly to github.com/settings/copilot/features) and disable the interaction data toggle. The process takes about 30 seconds — do it now rather than risk forgetting before the April 24 deadline.