As of April 16, 2026, Atlassian began rolling out a data contribution policy that, by default, feeds customer content from Jira, Confluence, and Jira Service Management into its AI training pipeline. Enforcement starts August 17, 2026. Depending on your plan tier, some of that data cannot be withheld — but if you’re on Free or Standard, your in-app content is already opted in and you can still turn it off.
What Atlassian Changed and Why It Matters Now
The policy affects organizations across Atlassian Cloud1. Starting August 17, Atlassian will use customer data to train Rovo AI, Rovo Chat, and automated agents — unless admins explicitly disable collection before that date.
The settings UI began rolling out April 16 and will reach all organizations by May 19, 20262. That’s an important caveat: if your organization hasn’t received the toggle yet, you are currently opted in with no mechanism to reverse it. The opt-out interface will arrive by May 19; enforcement doesn’t begin until August 17, so there is time — but the window is fixed.
What Data Gets Collected
Atlassian collects two distinct categories of data1:
Metadata includes readability scores, task classifications, semantic similarity scores, story points, sprint end dates, and SLA values. This is derived signal — computed from your content rather than the content itself.
In-app data includes Confluence page titles and body text, Jira work item titles, descriptions, and comments, plus custom emoji, status names, and workflow names.
The distinction matters because metadata and in-app data have different opt-out rules depending on your plan. Atlassian states that collected data is de-identified and aggregated before use in training4, and retained for up to seven years.
One point worth noting: “de-identified” metadata is not the same as harmless metadata. Semantic similarity scores derived from your Jira issues can, in aggregate, reveal product topology, team structure, and roadmap sequencing — information that competitive intelligence teams actively seek. The de-identification claim mitigates some risk but does not eliminate it.
The Tier Matrix: What Your Plan Actually Allows You to Do
Your plan tier determines which levers you can pull2:
| Plan | Metadata opt-out | In-app data default | In-app data opt-out |
|---|---|---|---|
| Free | Not available | ON | Available |
| Standard | Not available | ON | Available |
| Premium | Not available | OFF | Available |
| Enterprise | Available | OFF | Available |
Free and Standard customers face the starkest exposure: metadata collection is mandatory with no opt-out, and in-app data is on by default. Disabling in-app data is the only lever available, and it should be treated as mandatory if your Jira or Confluence instance contains anything proprietary.
Premium customers are in a better position — in-app data is off by default — but metadata collection remains non-negotiable.
Enterprise customers have the most control: both categories are off by default, and full opt-out of both is available. Note that some coverage has incorrectly characterized Enterprise as unable to opt out of metadata; Atlassian’s own support documentation confirms Enterprise can disable both2.
Step-by-Step: The Exact Settings Path to Opt Out
The settings live in one place2:
- Go to
admin.atlassian.com - Select Security in the left navigation
- Click Data contribution
- Toggle off In-app data (and Metadata, if you are on Enterprise)
Organization-level admin access is required. This is not accessible from within individual Jira or Confluence product settings — it must be done from the Atlassian Administration console.
Granular Control: Excluding Specific Spaces and Projects
If a full opt-out is not your preference, Atlassian provides more surgical controls2. From the same Data contribution panel, admins can exclude:
- Specific Confluence spaces
- Specific Jira projects or apps
- Individual Teamwork Graph connectors
This allows organizations to permit training on lower-sensitivity content (public-facing documentation, for example) while protecting internal roadmaps, security runbooks, or HR-adjacent Confluence spaces.
Who Is Automatically Exempt
Certain organization configurations are excluded from data contribution entirely — the policy cannot be turned on for them5:
- Organizations using customer-managed encryption keys (CMK/BYOK)
- Atlassian Government Cloud customers
- Atlassian Isolated Cloud customers
- Organizations subject to HIPAA compliance requirements
If your organization falls into any of these categories, no action is required. Data contribution is off by default and the toggle to enable it is not available.
GDPR and Enterprise Compliance Considerations
Atlassian’s position, as stated in its FAQ, is that the data contribution policy does not change its GDPR compliance obligations and that it continues to operate under its existing Data Processing Addendum and Privacy Policy5.
That is Atlassian’s legal position — not a legal conclusion that EU-regulated organizations can rely on without independent assessment. GDPR Articles 13 and 14 impose transparency obligations about processing purposes; if training AI models constitutes a new processing purpose not disclosed in existing privacy notices, organizations may face their own disclosure obligations to data subjects. The purpose limitation principle (Article 5(1)(b)) may require separate analysis depending on how Atlassian’s DPA characterizes the scope of permitted processing.
Organizations subject to GDPR, especially those that have signed Data Processing Agreements with Atlassian, should have their legal or compliance teams review whether the new policy triggers any re-notification obligations before August 17.
Atlassian is hosting an information webinar on April 28, 2026 that may address compliance questions in more detail3.
Deadline, Retention, and What Happens After You Opt Out
The key dates2:
- April 16 – May 19, 2026: Settings UI rolling out to all organizations
- August 17, 2026: Enforcement begins — data usage reflects whatever settings are configured at that point
- After opting out: In-app data is removed within 30 days; metadata within 90 days; affected models are retrained
Retention for data that was collected before an opt-out runs up to seven years4. Opting out stops future collection and triggers removal of previously collected data on the timelines above — it does not have retroactive legal effect on data already processed into model weights before the opt-out was registered.
FAQ
My organization is on Premium. Do I need to do anything?
In-app data is off by default for Premium, so your Confluence and Jira content is not being contributed unless someone has explicitly enabled it. Metadata collection is on and cannot be disabled on Premium. Log in to admin.atlassian.com > Security > Data contribution to confirm the in-app data toggle is in the expected state.
If I opt out now, does it stop retroactively? No. Atlassian will remove collected in-app data within 30 days of opt-out and metadata within 90 days, and will retrain affected models2. Data already processed into model weights before your opt-out is not recalled — opting out governs future collection and stored raw data, not model parameters already updated.
Does this affect Atlassian Data Center or Server customers? The policy applies to Atlassian Cloud products only. Data Center and Server deployments are self-hosted and not subject to this collection mechanism.
Can individual team members opt out, or does it have to be an admin? This is an organization-level control requiring Atlassian Administration access. Individual users cannot opt themselves out — the setting governs the entire organization’s data contribution posture2.
Footnotes
-
Data contribution settings | Atlassian Support ↩ ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8
-
Atlassian Enables Default Data Collection to Train AI | Hacker News ↩ ↩2
-
Atlassian Will Collect Jira and Confluence Data by Default to Train AI Models – gHacks Tech News ↩ ↩2
-
Data Contributions Frequently Asked Questions | Atlassian ↩ ↩2