Amid the AI ​​frenzy, fear and anger grow in the meta

AI For Business


Meta employees have weathered frequent layoffs since early 2025, including this spring when the company cut 10% of its workforce. (AFP photo)
san francisco:

The frenzied push for artificial intelligence dominance comes at a different kind of price for Meta. Mass layoffs, employee monitoring and turnover have fueled reports of a heated internal environment.

As Meta spends billions of dollars each year to build AI capabilities, employees at Facebook, Instagram and WhatsApp are increasingly frustrated with Mark Zuckerberg’s parent company.

Meta employees have endured frequent layoffs since early 2025. That includes the company cutting 10% of its workforce (about 8,000 people) this spring and reassigning an additional 7,000 employees.

For those who remain, internal AI training plans have drawn accusations of surveillance.

The company also underwent a major restructuring of its AI research division, into which Zuckerberg, Meta’s founder and chief executive, has poured billions of dollars.

This slump is in stark contrast to Meta’s strong financials, which are supported by advertising, which accounts for nearly 98% of its revenue. In the first three months of 2026, Meta’s net profit increased to more than USD 26 billion.

But the bill for AI investments has also skyrocketed, leading Mr. Zuckerberg, who has near-absolute power over the company, to impose drastic staff cuts and increased oversight in the name of efficiency and savings.

These cuts are funding a massive infrastructure race. Meta plans to spend up to US$145 billion on AI investments this year, nearly twice as much as last year.

Data collection

After thousands of employees were redeployed to Meta’s AI division, some employees who spoke anonymously to US media complained of the “daunting” tasks designed to train machines and even automate their jobs.

The controversial program, called the Model Capabilities Initiative, was launched in April and suspended on June 22nd. The program captured the clicks, keystrokes, and browsing activities of U.S. employees to train an AI agent, a piece of software that can perform tasks independently.

Zuckerberg, who has made AI his company’s north star, defended the program in an internal meeting, saying, “AI models learn by watching really smart people do things,” Wired reported.

However, this tool caused a revolt. More than 1,600 employees have signed a petition calling for its repeal, with some likening the company to a “data extraction factory,” according to media reports.

The suspension occurred after private conversations and performance data were accidentally made accessible to all staff. The system risked attracting the attention of European regulators because it recorded communications between employees on both continents.

A Meta spokesperson said in a statement to AFP on Tuesday that the program is designed to protect privacy.

“At this time, there is no indication that Meta employees have compromised data, but we have suspended access to the data while we investigate,” the statement said.

One employee summed up the mood with a meme from “The Office” posted on a company forum that read, “0 days since last nonsense.”

“Dead End Quest”

All of these efforts are aimed at making up for the persistent lag to Google, OpenAI, and Anthropic, which dominate the race for cutting-edge AI models. Meta’s own model suffered from repeated delays and proved disappointing internally.

To regain ground, Mr. Zuckerberg last year invested more than $14 billion in San Francisco-based startup Scale AI and poached CEO Alexander Wang, then 28, to run a “superintelligence” lab within Meta.

This high stakes hasn’t caught people’s attention yet. Several key figures have since stepped down, including Yann LeCun, considered one of the “godfathers” of modern AI and who has led Meta’s AI research since 2013.

Mr. LeCun suddenly found himself subordinate to Mr. Wang, who was more than 35 years his junior. He left Meta at the end of 2025 and launched his own startup.

In an interview with the Financial Times, the Turing Award winner lamented that although Wang is a “quick learner”, he has “no research experience” and pursues “dead ends”.

For Meta, the stakes now go beyond social networks. The company is also focusing on home appliances with smart glasses, and is considering a new market forecasting app called Arena, potentially in partnership with Polymarket and Calci, according to the New York Times.

Litigation can also consume time and resources.

A Los Angeles jury in March found Mehta responsible for the effects of his social media addiction for the first time, a day after another ruling in New Mexico that found him failing to protect a minor.

Meta has appealed, and more lawsuits are expected this year.



Source link