startups: Meta employees protest new mouse-tracking software days

Meta employees across several US offices have launched a visible protest against the company's new mouse-tracking software, the “Model Capability Initiative,” just days before widespread layoffs are set to impact 8,000 staff members. Flyers distributed throughout offices Tuesday denounced the program as an “Employee Data Extraction Factory,” fueling concerns over workplace surveillance and job automation under the guise of AI training.

The initiative, which captures mouse movements, clicks, keystrokes, and screenshots on a specific list of work applications, has been met with significant internal pushback. Employees, citing the National Labor Relations Act and their right to organize, are being urged to sign an online petition against the software. Concurrently, UK-based Meta colleagues have initiated a unionization drive with United Tech and Allied Workers, further amplifying the dissent.

Meta asserts that the data collected through the Model Capability Initiative is crucial for teaching AI agents how humans interact with software, enabling them to complete everyday tasks. The company stated that such models require “real examples of how people actually use them,” including mouse movements and navigation. It also highlighted that safeguards are in place to protect sensitive company information and that the program operates only on designated applications and websites, not across all computer activity.

However, many employees, as reported by Reuters, interpret the program as a form of workplace surveillance disguised as training data. This perception is particularly acute given its proximity to Meta’s May 20 layoffs, which will reduce the workforce by approximately 10 percent. Further job cuts are anticipated in the latter half of 2026. Chief Executive Mark Zuckerberg's January statement that 2026 would be “the year that AI starts to dramatically change the way that we work” is now being read internally as a direct reference to jobs being converted into datasets for automation.

While Meta describes the Model Capability Initiative as voluntary in spirit, it is effectively mandatory for staff utilizing the designated applications on company-issued machines. The legality and ethical implications of such a program remain ambiguous, especially in jurisdictions with stricter employee privacy regulations, such as the EU, which demands a higher bar for proportionality and worker consent than current US federal law.

This protest marks a rare instance of organized internal dissent at Meta, a company that has historically managed to avoid widespread staff unrest. A previous visible bout of employee dissatisfaction in 2018 regarding sexual-harassment policies ultimately led to policy changes. However, this current situation is compounded by earlier challenges, including a data breach at Meta’s AI data-labelling vendor, Mercor, last month, which temporarily halted some data work.

As the May 20 layoffs loom, the online petition against the Model Capability Initiative continues to circulate, underscoring the ongoing tension between Meta’s AI ambitions and its employees’ concerns over privacy and job security.

FAQ

Q: What is the Model Capability Initiative?

A: The Model Capability Initiative is new software implemented by Meta that captures mouse movements, clicks, keystrokes, and screenshots on a designated list of work applications used on company-issued machines. Meta states its purpose is to train AI agents on how humans navigate software.

Q: Why are Meta employees protesting this software?

A: Employees view the software as a form of workplace surveillance and a step towards automating their own jobs, particularly concerning given the impending mass layoffs. They have dubbed it an “Employee Data Extraction Factory” and are organizing a petition and unionization drive against it.

Q: What are the potential implications of this software, especially in the context of the upcoming layoffs?

A: Employees fear the data collected could directly contribute to the automation of their roles, aligning with CEO Mark Zuckerberg's statement about AI changing how work is done. The software also raises significant privacy concerns, particularly regarding its voluntary-yet-mandatory nature and its compliance with stricter employee privacy laws in regions like the EU.