Meta pauses employee surveillance for AI training: What it means for your privacy

In late June 2026, Meta temporarily halted an internal tool that tracked employee productivity — and used that data to train artificial intelligence models. The move came after backlash from staff and privacy advocates, and it highlights a growing practice: companies collecting personal data from any available source to feed AI systems, often without clear consent.

This isn’t just a workplace issue. If your data flows through a company that builds AI, there’s a chance it’s being repurposed. Here’s what happened, why it matters for anyone who uses Meta services or works in a data-rich environment, and what you can do about it.

What happened

According to a Guardian report published June 25, 2026, Meta had been running a tool that monitored employee activity — including keystrokes and app usage — and used that data to train its AI models. After employees raised concerns and privacy advocates criticized the lack of transparency, Meta paused the program.

Meta did not say whether the tracker would be permanently discontinued. The company has a long history of privacy controversies, from the Cambridge Analytica scandal to repeated fines by European regulators for data breaches. This latest episode fits a pattern: Meta has been aggressively reorganizing its workforce around AI, and internal data — even sensitive workplace data — is seen as a resource.

Why this matters for all users

The Meta case is a reminder that companies often treat any data they collect as fair game for AI training. It’s not limited to social media posts or search queries; it can include your behavior at work, your keystroke patterns, how long you spend in certain apps, and more.

Meta is not alone. Other tech giants have faced scrutiny for similar practices. Amazon has used warehouse worker tracking data to optimize shift scheduling and, reportedly, to train models that evaluate productivity. Google has faced questions about using data from its workplace tools (like Google Docs and Calendar) for AI improvements. In each case, the line between operational monitoring and AI training is blurry.

The deeper concern is consent. Did Meta employees know their keystrokes would be used to train an AI? Based on the Guardian’s reporting, many did not. For regular consumers, the same opacity applies: terms of service often permit companies to use your data for “improving services,” which increasingly means feeding AI models.

What readers can do

You can’t stop every company from repurposing your data, but you can reduce your exposure and push for better protections.

Check your employer’s data policies. If you work at a company that uses productivity tracking software — even standard tools like Microsoft 365 or Slack — ask whether that data is used for AI training. Some employers have internal transparency reports; if not, consider requesting one through HR or worker representatives.

Adjust your Meta privacy settings. On Facebook and Instagram, go to Settings > Privacy > Data Use. Look for options to limit how your content is used for AI training. Meta has a “Data for AI” toggle in some regions, but it may not be available everywhere. Also review your ad preferences: AI models often learn from behavioral data collected for ad targeting.

Use opt-out tools where available. Some platforms offer forms to request that your data not be used for AI training (for example, OpenAI and Google have opt-out pages for certain services). These are not always honored immediately, but submitting them creates a paper trail.

Support stronger data protection laws. The European Union’s AI Act includes provisions for transparency and consent in AI training data. In the US, several states are considering similar bills. Writing to your representatives or supporting advocacy groups like the Electronic Frontier Foundation can help.

Consider privacy-focused alternatives. For communication and productivity, tools like Signal, ProtonMail, or standard encrypted messaging apps collect minimal data. For social networking, services that don’t rely on advertising revenue (such as Mastodon or Bluesky) have less incentive to harvest your activity for AI.

Sources

  • The Guardian, “Meta pauses employee tracker for AI training amid privacy concerns,” June 25, 2026.
  • The Guardian, “Meta is rapidly reorganizing its workers’ jobs around AI: ‘Transfers aren’t optional’,” May 19, 2026.
  • Previous reporting on Cambridge Analytica, EU fines against Meta, and Amazon workplace surveillance (for context; not linked here due to space, but widely documented).