Atlassian to Use Customer Data for AI Training Across Platform, Introduces Opt-Out Controls
Key Takeaways
- ▸Atlassian will begin using aggregated customer data across its platform to train AI and improve products for all users starting August 17, 2026
- ▸New in-app controls and settings will allow customers to opt out of data contribution, initially available for Jira, Confluence, and Jira Service Management
- ▸The company is implementing de-identification and aggregation safeguards to protect customer data while leveraging insights for AI model improvement
Summary
Atlassian announced a significant policy shift that will begin on August 17, 2026, allowing the company to use customer metadata and in-app data to train AI models and improve products across its entire customer base, rather than only for individual organizations. The change applies initially to Jira, Confluence, and Jira Service Management, along with Atlassian Platform apps including Rovo, Home, Teams, Projects, Assets, Goals, and Analytics.
The company is implementing new data contribution settings in Atlassian Administration to give customers control over whether their data is used for collective AI improvement, and strengthening privacy protections through de-identification and aggregation techniques. Atlassian framed the change as essential for delivering enhanced AI capabilities while maintaining transparency and data security, positioning it as a balanced approach between innovation and customer trust.
Currently, Atlassian uses metadata and in-app data only to improve individual customer experiences within their own instances. The policy expansion aims to leverage a broader, more diverse dataset to enhance AI features that benefit all customers, from enterprise search and chat to AI agents that automate workflows.
- The change represents a shift from using data only to improve individual customer instances to using anonymized, aggregated data for collective platform improvements
Editorial Opinion
Atlassian's decision to harness customer data for collective AI improvement reflects an industry trend of balancing innovation with customer privacy concerns. While the company's introduction of granular opt-out controls and privacy-preserving techniques is commendable, the August 2026 effective date raises questions about whether customers will have sufficient time to understand the changes and make informed decisions before data sharing begins. The initial limitation to Jira and Confluence products suggests more products will follow, and customers should carefully evaluate their data contribution settings as they become available.
