Meta tracking employee computer interactions for AI agent training

What Meta is doing

Meta is deploying internal tracking software on US-based employee computers to collect interaction data for AI model training. According to internal memos seen by Reuters, this is part of their initiative to build AI agents capable of performing work tasks autonomously.

Specific data being collected

• Mouse movements
• Clicks
• Keystrokes
• Occasional screen snapshots for context

How the tool works

The tracking software runs on a list of work-related applications and websites. According to the memo posted by a staff AI research scientist in Meta's internal channel for the Meta SuperIntelligence Labs team, the tool will capture these inputs on certain applications.

Purpose and use cases

Meta spokesperson Andy Stone stated: "If we're building agents to help people complete everyday tasks using computers, our models need real examples of how people actually use them - things like mouse movements, clicking buttons, and navigating dropdown menus."

The internal memo specifically mentioned improving models in areas where they struggle, including:

• Choosing from dropdown menus
• Using keyboard shortcuts

The memo stated: "This is where all Meta employees can help our models get better simply by doing their daily work."

Privacy and usage limitations

Meta spokesperson Andy Stone clarified that the collected data will not be used for:

• Performance assessments
• Any other purpose besides model training

Stone also mentioned that safeguards are in place to protect sensitive content.