Leaked Audio Reveals Meta Tracks Employee Keystrokes to Train AI as 8000 Layoffs Begin

A leaked audio recording from an internal Meta all-hands meeting has ignited a firestorm of controversy. The recording, which went viral on May 19, 2026, appears to capture CEO Mark Zuckerberg defending a mandatory employee surveillance program designed to harvest keystrokes, mouse movements, and screenshots for artificial intelligence training.

The leak arrives on the same morning that approximately 8,000 Meta employees received layoff notices, representing roughly 10 percent of the company’s global workforce.

What the Leaked Audio Reveals

The recording allegedly comes from an April 30, 2026, internal meeting where an employee directly questioned Zuckerberg about tracking software installed on US-based company computers. The employee stated, “I know layoffs are top of mind, but there were also some updates this week around employee device tracking,” before asking the CEO to explain the monitoring.

Zuckerberg responded by arguing that modern AI systems improve by observing highly skilled people performing real-world tasks. He explained that Meta believes its own engineers provide superior training data compared to outside contractors. “The average intelligence of the people who are at this company is significantly higher than the average set of people you can get to do tasks,” Zuckerberg allegedly said in the recording.

He specifically noted that teaching AI models to code using internal engineers would dramatically improve the company’s coding capabilities faster than competitors can match.

The Model Capability Initiative Program

The surveillance program, known internally as the Model Capability Initiative or MCI, was first reported in April 2026. The software runs on approved applications, including Gmail, GChat, VSCode, and Meta’s internal AI assistant Metamate. It captures mouse clicks, keystrokes, and occasional screenshots of employee devices.

Meta spokesperson Andy Stone previously stated that the data helps AI models understand everyday computer actions such as clicking buttons, navigating menus, and using software interfaces. The company emphasized that the collected data will not be used to evaluate employee performance.

No Opt-Out Option for Employees

When employees asked whether they could avoid the surveillance, Meta Chief Technology Officer Andrew Bosworth delivered a blunt response. According to the leaked meeting details, Bosworth stated that “there is no way to opt out if you use a company computer” . The response reportedly received over a hundred shocked emoji reactions from employees.

The program currently applies only to US based workers. Implementing similar monitoring in Europe would face significant obstacles due to strict GDPR data protection laws. Yale Law Professor Ifeoma Ajunwa noted that “there is no limit on worker surveillance” in the United States at the federal level.

Employee Backlash and Internal Resistance

The controversy has sparked rare public dissent within Meta. One engineer who worked at the company for 11 years posted an internal manifesto that was viewed by nearly 20,000 colleagues. The engineer wrote, “I do not want to live in a world where humans are exploited as training data for AI” .

An internal petition circulating among employees demands that Meta stop non-consensual data collection and respect employee privacy. The petition specifically states that “companies should not extract employee data for AI training without consent”.

Some employees have begun delaying installation of the monitoring software despite persistent system reminders. Others have posted反对 signs in offices and restrooms, though Meta reportedly removes some while bathroom posters often survive longer.

Timing Raises Suspicions

The leak comes as Meta executes a massive strategic shift away from元宇宙 toward artificial intelligence. Beyond the 8,000 layoffs announced on May 20, the company has reportedly reassigned over 7,000 employees to AI-focused roles. These employees are being directed to help develop AI systems that could eventually automate the very tasks they currently perform.

One employee summarized the situation bluntly. The company demands full use of AI tools, tracks computer activity to feed models, measures AI usage in performance reviews, and then replaces workers with AI.

Broader Implications

The controversy highlights growing tensions across the technology industry as companies race to dominate artificial intelligence. Critics argue that using employee monitoring for AI training blurs the line between productivity tracking and the extraction of intellectual labor as raw material for automation.

AI expert and AIEnsured Chief Technology Officer Dr Srinivas Padmanabhuni told NDTV that this issue reflects a larger pattern across Big Tech where user behavior and digital activity have increasingly become valuable training material for AI models.

For Meta employees, the message is clear. The software running quietly in the background may no longer just monitor productivity. It may be learning how to replace you

Scroll to Top