GitHub to use Copilot interaction data from individual users for AI training unless they opt out
The Facts
- GitHub announced that, beginning April 24, it will use Copilot interaction data from Free, Pro, and Pro+ users to train and improve its AI models unless those users opt out.
- The policy applies to individual Copilot tiers including Free, Pro, and Pro+, while Copilot Business and Copilot Enterprise accounts are excluded.
- GitHub describes the training data as including Copilot inputs, outputs, code snippets, and associated context.
- Reports on the policy say the collected interaction data may also include items such as chats or feedback, comments or documentation, file names, repository structure, and navigation patterns tied to Copilot use.
- GitHub says users can opt out through Copilot privacy settings, and users who had already opted out do not need to take further action.
- GitHub says it wants to use real-world Copilot interaction data to better understand development workflows and improve the quality, safety, and relevance of its coding assistance.
- The change matters for developers and organizations because it ties routine Copilot use to model-training data, prompting renewed privacy, compliance, and data-governance questions about how coding interactions are handled.
- One unresolved issue for affected users is whether to remain in the default training program or opt out before or after the policy takes effect, especially for those using Copilot on personal or non-enterprise projects.
Context
Which GitHub Copilot users are affected?
GitHub says the update affects Copilot Free, Pro, and Pro+ users. Copilot Business and Copilot Enterprise are excluded, and multiple reports say student and teacher access is also exempt GitHub Blog,heise online,GEO TV.
What kinds of data may be used for training?
GitHub says training data includes Copilot inputs, outputs, code snippets, and associated context. Other reports describing the policy say this can also include items such as comments, file names, repository structure, chats, feedback, and navigation patterns connected to Copilot interactions GitHub Blog,TechSpot,blockchain.news.
How can users stop their Copilot interaction data from being used for training?
GitHub says affected users can opt out in Copilot settings under the Privacy section. It also says users who previously opted out will keep that preference and do not need to act again GitHub Blog,TheRegister.com,IT Security News - ….
How left and right read it
What stands out is how much ordinary developer activity can now flow into model training by default: inputs, outputs, code snippets, associated context, and reported items like chats, comments, file names, repository structure, and navigation patterns tied to Copilot use. Even with an opt-out, linking routine use to training raises real privacy, compliance, and data-governance concerns for people working on personal or non-enterprise projects.
The key issue is user choice and clear boundaries. GitHub is limiting this policy to individual Copilot tiers while excluding Business and Enterprise accounts, and it says users can opt out through privacy settings, but affected users still have to decide whether to stay in the default training program as the company uses real-world interaction data to improve quality, safety, and relevance.
The receipts — 21 sources
Independent coverage (21)
The news, read by both sides.
The day’s biggest stories in one short brief, built to show both sides and the competing values behind the headlines. Free in your inbox.