Meta‘s 8,000-person layoff round began on Wednesday, with notification emails sent to affected employees globally in three batches, each at 4 a.m. local time — a rolling sequence that started in Asia and ended on the US West Coast.
Alongside the cuts, a leaked audio clip from an internal all-hands meeting has gone viral, and an employee revolt against the company’s AI training program is intensifying.
Read: Meta Announces 8,000 Layoffs on May 20 as AI Investment Surges
LEAKED AUDIO FROM META ALL-HANDS AHEAD OF LAYOFFS TOMORROW
— Official Layoff (@LayoffAI) May 19, 2026
Mark Zuckerberg, in his own words, told Meta employees their devices are being tracked to train AI models.
His reasoning? Meta employees are smarter than the contract workers the rest of the industry uses for data… https://t.co/VSPdjHZ2ga pic.twitter.com/3TX0vLP8P3
In late April, Meta quietly disclosed the Model Capability Initiative (MCI) — a tracking program installed on US-based employees’ work computers that captures keystrokes, mouse movements, clicks, and periodic screenshots. The tool runs across hundreds of apps and websites, including Google, LinkedIn, GitHub, Slack, and Meta’s own platforms.
Read: Meta Starts Tracking Employees to Train AI
Meta says the data will train AI agents to handle routine office tasks — the kind of work its employees do every day. An internal memo from Meta Superintelligence Labs put it plainly: the company needs a “big and unbiased” dataset of how people actually use computers. In other words, the workers losing their jobs this week are also the ones teaching the models that may replace them.
Meta spokesperson Andy Stone said the data would be used solely for model training, not performance reviews, and that personal content would be stripped from training sets before use. European employees are exempt, as the program conflicts with GDPR. No equivalent protection exists for US-based staff.
Workers launched a petition urging Zuckerberg to shut the program down. “Collecting and repurposing this kind of data raises serious concerns around privacy, consent, and trust in the workplace,” it read. “It should not be the norm that companies of any size are permitted to exploit their employees by nonconsensually extracting their data for the purposes of AI training.”
Meta CTO Andrew Bosworth’s assurances that the data would be “tightly controlled” have done little to quiet the unrest. “There’s an emerging sense of dread across wide swaths of the company,” a current Meta employee told CNBC.
The leaked audio clip purportedly from a Meta all-hands meeting on April 30 has circulated widely on social media, with accounts attributing specific quotes to Zuckerberg about employee intelligence and the MCI program’s rationale.
Zuckerberg has said publicly that 2026 will be “the year that AI dramatically changes the way we work,” and that MCI forms part of a broader strategy to build AI agents capable of eventually replacing the human labor currently performing routine cognitive tasks.
Meta is spending up to $145 billion on AI infrastructure this year — a figure the company raised from its original $115–$135 billion guidance after higher component costs. It is simultaneously cutting the workforce that generates the data feeding those models, freezing 6,000 open roles, and redirecting roughly 7,000 employees into AI-focused positions.
Chief People Officer Janelle Gale told staff at an internal meeting that she could not guarantee there would be no further layoffs. “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,” Gale told employees, according to Business Insider, which cited three people on the call.
Information for this story was found via the sources and companies mentioned. The author has no securities or affiliations related to the organizations discussed. Not a recommendation to buy or sell. Always do additional research and consult a professional before purchasing a security. The author holds no licenses.