GitHub Copilot Will Train on Your Code Starting April 24 — Here's How to Opt Out
GitHub quietly dropped a policy update that's worth every developer's attention: starting April 24, 2026, interaction data from Copilot Free, Pro, and Pro+ users will be used to train AI models by default. That means inputs, outputs, code snippets, and session context could all be feeding future versions of Copilot unless you explicitly opt out. Business and Enterprise customers are excluded, and anyone who previously opted out is already covered — but for the tens of millions of individual Copilot users, a manual settings change is now required.
GitHub's stated rationale is straightforward: real-world coding data helps the model understand actual developer workflows, catch subtle bugs, and suggest more secure patterns. There's genuine merit to that argument — models trained on sanitized or synthetic data have real limitations. But the opt-out-by-default framing landed poorly with the developer community. On Hacker News, the post hit 312 points and generated 146 comments, with the bulk of the pushback centered on the lack of prominent notification and the passive enrollment of users who may not regularly check policy updates.
If you're on a personal Copilot plan, the fix takes about 30 seconds: navigate to github.com/settings/copilot/features, find the Privacy section, and toggle off "Allow GitHub to use my data for AI model training." You have until April 24 to make the change before the policy takes effect.