Atlassian's AI Data Collection Plan: Pay More or Opt Out (Maybe)

Atlassian is rolling out a controversial new data collection policy that will see the company gather user metadata and in-app content from its cloud products to train AI models, unless customers pay for the most expensive Enterprise tier or local laws intervene.

What Data Is Being Collected?

Starting August 17, 2026, Atlassian will collect two categories of data from its 300,000 global customers:

Metadata Collection (Always On for Lower Tiers)

• Readability scores and complexity ratings for Confluence page content
• Task classifications assigned to content (e.g., "sales work item")
• Semantic similarity scores measuring how similar two Confluence pages are
• Numbers entered into Atlassian-created fields (story points, sprint end dates, SLA requests)

In-App Data Collection

• Titles and content in Confluence pages
• Titles, descriptions, and comments in Jira work items
• Custom emoji names, custom status names, and custom workflow names

The Pay-to-Privacy Model

The policy creates a clear tiered approach to data privacy:

Free, Standard, and Premium tiers: Metadata collection is mandatory and cannot be opted out of. In-app data collection is on by default but can be disabled.

Enterprise tier: Both metadata and in-app data collection are turned off by default, giving these customers more control over their data.

Who Gets Exemptions?

Several customer categories are completely excluded from data collection:

• Customers using customer-managed keys (bring your own key)
• Atlassian Government Cloud users
• Atlassian Isolated Cloud users
• Customers with HIPAA compliance requirements
• Some government and financial services customers

How Long Will Data Be Kept?

Atlassian plans to store collected data for up to seven years. The company states this extended retention enables "deeper insights into customer behavior" and helps drive continual improvements to the overall user experience.

What Happens When You Opt Out?

For customers who can opt out (Premium tier and below for in-app data):

• Data is removed from Atlassian's datasets within 30 days of opting out or deleting apps
• Models previously trained on that data will be re-trained within 90 days

The AI Training Angle

Atlassian claims the collected data will help its AI models:

• Surface more relevant results in response to prompts and queries
• Summarize content more accurately and concisely
• Identify the best templates for creating new documents
• Learn which agentic workflows and follow-up questions lead to successful completions

Industry Context

This move aligns with a broader industry trend where SaaS companies are increasingly using customer data to train AI models. However, Atlassian's approach of making data collection mandatory for lower-tier customers while offering privacy as a premium feature has drawn criticism for creating a two-tiered privacy system.

Compliance and Legal Considerations

The policy explicitly states that data collection will not occur where "the law forbids it," suggesting Atlassian is aware of potential regulatory hurdles, particularly in regions with strict data protection laws like the EU's GDPR.

Timeline and Implementation

The new data collection settings will not apply to customers who terminate their contracts before August 17, 2026. After that date, the policy becomes enforceable for all active customers.

This policy represents a significant shift in how Atlassian handles customer data and raises important questions about privacy, consent, and the commercialization of user-generated content in the age of AI. Enterprise customers and those with specific compliance requirements will have more control over their data, while smaller organizations and individual users will have their data harvested by default unless they can afford to pay more or are protected by law.