GitHub Copilot Interaction Data Usage Policy Update

GitHub will use interaction data from Copilot Free, Pro, and Pro+ users for AI model training and improvement starting April 24. This includes inputs, outputs, code snippets, and related context. However, Copilot Business and Enterprise users are not affected by this update.

Users can opt out of data collection at any time via the 'Privacy' menu in settings. For users who have already set an opt-out, that setting will remain in effect and their data will not be used for training.

Through this data usage, models can better understand development workflows, suggest more accurate and secure code patterns, and improve their ability to detect bugs in advance. GitHub has already introduced interaction data from Microsoft employees into training and confirmed cases of significant model performance improvement.

Data collected:

• Outputs accepted or modified by users
• Inputs sent to GitHub Copilot (including code snippets)
• Code context around the cursor position
• Comments and documentation written
• File names, repository structure, and navigation patterns
• Interactions and feedback with Copilot features (Chat, inline suggestions, etc.)

Data excluded from collection:

• Data from Copilot Business, Enterprise, or enterprise-owned repositories
• Data from users who have opted out of model training in settings
• Issues, discussions, or private repository content at rest (however, processing of private repository code that occurs during Copilot use is necessary for service execution and may be used for training if not opted out)