Meta Confirms Employee Tracking to Train AI Models, Cites Intelligence Advantage

Meta Platforms has confirmed it is tracking the computer activity of its US-based employees to collect data for training artificial intelligence models. CEO Mark Zuckerberg addressed the move in a company-wide meeting, stating that Meta employees provide higher-quality data than external contractors. This follows reports that Meta installed software to monitor mouse movements, clicks, keystrokes, and take occasional screenshots on work-related apps and websites.

Key Highlights

• Meta tracks US employees' computer activity to collect data for AI model training.
• CEO Mark Zuckerberg says internal data is higher quality than contractor-generated data.
• Meta's Model Capability Initiative monitors mouse movements, keystrokes, and screen content.
• Company plans to lay off about 10 percent of staff next month, with more cuts possible.
• AI automation is not the main driver of layoffs, according to Meta leadership.

Meta’s Data Collection Strategy

The internal tracking tool, called Model Capability Initiative (MCI), aims to gather detailed information on how employees interact with computers. The company will use this data to replicate human behaviors, such as selecting options from dropdown menus and using keyboard shortcuts. According to Zuckerberg, Meta believes its employees are significantly more intelligent than the average contractor used by other AI firms. He said this approach could give Meta an edge over competitors that rely on external data sources.

Meta CTO Andrew Bosworth previously told staff that the company would increase internal data collection as part of its “AI for Work” efforts. This initiative has since been renamed Agent Transformation Accelerator (ATA). The company’s focus is on using high-quality internal data to improve AI performance and efficiency.

Layoff Plans and Company Outlook

During the same meeting, Meta’s chief people officer Janelle Gale discussed upcoming layoffs. She said the company plans to reduce its workforce by about 10 percent next month. Gale did not rule out further cuts, emphasizing the need to manage costs amid changing priorities and strong competition. She stated, “Will there be more layoffs? The question always comes up. I’d love to say that there are no more layoffs, but I can’t say something we can’t deliver.”

Zuckerberg also addressed the topic of layoffs, clarifying that AI automation is not the main reason for the staff reductions. He noted that AI has made small teams more efficient, but the layoffs are driven by broader business considerations.

Industry Context

Meta’s approach to AI training differs from many competitors, who often use third-party contractors to generate data. By leveraging its own workforce, Meta aims to improve the quality of data used to train its AI models. The company’s strategy reflects a broader trend in the tech industry to prioritize efficiency and innovation while managing operational costs.