SUBSCRIBE TO OUR FREE NEWSLETTER
Daily news & progressive opinion—funded by the people, not the corporations—delivered straight to your inbox.
5
#000000
#FFFFFF
5
#000000
#FFFFFF
In Leaked Audio, Zuckerberg Tells Meta Workers He's Been Using Them to Train AI Ahead of Mass Layoffs
"AI is a freight train, but the future is not a foregone conclusion," said one engineer, urging his colleagues to sign a petition to stop Meta's use of an AI tracking program. "It’s not too late to pump the brakes."
May 20, 2026
Meta employees reported Wednesday that in the company's offices on the day mass layoffs hit thousands of their colleagues, fliers were taped to walls urging workers to sign a petition in support of stopping the company's new artificial intelligence data tracking program—which CEO Mark Zuckerberg touted late last month as a way for its new AI models to "learn from watching really smart people do things."
A day before about 8,000 Meta employees began receiving emails notifying them that they were being laid off—a process that began in Singapore at 4:00 am local time Wednesday and continued in European and US offices in their respective time zones—the labor-focused media organization More Perfect Union shared a leaked audio file in which Zuckerberg was heard explaining how the AI training program worked.
"The average intelligence of the people who are at this company is significantly higher than the average set of people that you can get to do tasks," said Zuckerberg. "So if we're trying to teach the models coding, for example, then having people internally build tools or solve tasks that help teach the model how to code, we think is going to dramatically increase our model's coding ability faster than what others in the industry have the capability to do, who don't have thousands and thousands of extremely strong engineers at their company."
LEAKED AUDIO: In an all-hands meeting on April 30, Mark Zuckerberg tells employees that he's training AI on them ahead of mass layoffs.
"The AI models learn from watching really smart people do things... The average intelligence of the people who are at this company is… pic.twitter.com/lt9eeJ3cwh
— More Perfect Union (@MorePerfectUS) May 19, 2026
He assured the company's 78,000 employees that "no human is looking at or watching what people are doing on their computers... None of the data is being used for looking at what people are doing or surveillance or performance tracking or anything like that. It's purely just that we are using this to feed a very large amount of content into the AI model so that way it can learn how smart people use computers to accomplish tasks."
Zuckerberg explained how the employees have been used to train the model that could potentially replace many of them days after Meta announced it was planning to lay off about 10% of its workforce as the company invests heavily in AI, spending $125 billion to $145 billion on the technology—more than double what it spent last year.
The New York Times reported earlier this month that employees "revolted" when they learned about the AI tracking program, and expressed fears that they had unknowingly been training a model that would ultimately replace them.
An engineering manager asked on the company's internal communication platform how workers can opt out of having their computer activity monitored to train the AI model, only to be told by chief technology officer Andrew Bosworth, "There is no option to opt out on your corporate laptop."
Another employee told Bosworth, “Your callousness to the concerns of your own employees is concerning."
On Monday, The New York Times reported, employees learned that in addition to the layoffs, another 7,000 workers will be reassigned to help develop AI tools.
About 2,000 employees began working this month on a new Applied AI and Engineering team, which is set to use the data gathered by the AI tracking program Zuckerberg described to build AI tools. Those who volunteered to join the group would not be included in this week's layoffs, the Times reported.
"Every company is training AI on their employees," said Chen Avnery, an independent adviser on AI governance and data platforms. "Meta just said it out loud. The question stopped being, 'Will AI replace you?' a year ago. Now it's whether you're building the agents or generating their training data."
More than 1,000 people in the company have signed the petition calling to halt the AI data program, according to the newspaper.
Software engineer Mack Ward urged his colleagues to sign on earlier this month, telling them in an internal post that "AI is a freight train, but the future is not a foregone conclusion."
"It’s not too late to pump the brakes and consider how we, society, want to go about this,” Ward said. “Speaking up is never easy, but ‘easy’ isn’t what you were hired to do.”
Keep ReadingShow Less
'Colossally Wrong Decision' as Facebook Parent Company Wins Instagram-WhatsApp Antitrust Case
"This court has effectively told every aspiring monopolist that our current justice system is on their side."
Nov 19, 2025
Anti-monopoly advocates are warning that a federal judge's ruling in favor of Facebook parent company Meta in a major antitrust case will have negative repercussions for US consumers by allowing Facebook to continue wielding monopoly power in the social media marketplace.
Judge James Boasberg in the District Court for the District of Columbia ruled Tuesday that the company’s acquisitions of Instagram and WhatsApp did not violate US antitrust policy.
Boasberg found that the Federal Trade Commission (FTC) had not proven Meta holds monopoly power in the personal social networking market, "largely because he folded TikTok and YouTube into the same market and concluded that their popularity reduces Meta’s share below illegal levels," said the American Economic Liberties Project (ALEP).
John Bergmayer, legal director at Public Knowledge, argued that Boasberg's ruling demonstrates a basic misunderstanding about the economics of the social media market.
"The court's opinion reflects a view of the market that is at odds with how digital-platform power operates today," he said. "Meta systematically acquired emerging competitors precisely because direct, head-to-head competition threatened its dominance. Meta’s consolidation strategy deprived consumers of innovative services and prevented the development of a truly competitive social-networking ecosystem."
Nidhi Hegde, executive director of ALEP, described the ruling as a "colossally wrong decision" that "turns a willful blind eye to Meta’s enormous power over social media and the harms that flow from it."
"These deals let Meta fuse Facebook, Instagram, and WhatsApp into one machine that poisons our children and discourse, bullies publishers and advertisers, and destroys the possibility of healthy online connections with friends and family," she said. "By pretending that TikTok’s rise wipes away over a decade of illegal conduct, this court has effectively told every aspiring monopolist that our current justice system is on their side."
Hegde added that it should now fall upon US Congress to "step in and break up Big Tech, prohibit addictive surveillance algorithms, and create the conditions for building a better future."
Open Markets Institute policy counsel Tara Pincock said Boasberg's ruling was "profoundly misguided," and accused the judge of blocking the FTC from reversing a mistake it made last decade when it signed off on Meta's purchases of Instagram and WhatsApp.
"Judge Boasberg erred in concluding that Facebook competes with TikTok and YouTube," said Pincock, a former state assistant attorney general in Utah. "I was part of the bipartisan coalition of states that brought this case alongside the FTC in December 2020, and the court’s framing misrepresents what is at stake. This case has never been about generic 'time and attention.' It is about how people connect, communicate, and build communities—and about how a powerful company abused its dominance to protect itself from competition."
Keep ReadingShow Less
Siding With Trump, Facebook Suspends Chicago Community Group Issuing ICE Alerts
"This is the same company that had policies allowing for chatbots to have 'sensual' conversations with kids, but discussing local law enforcement among neighbors is a bridge too far, huh?"
Oct 15, 2025
Chicago residents in recent weeks have found numerous ways to resist the Trump administration's deployment of hundreds of federal agents in its increasingly violent "Operation Midway Blitz" anti-immigration campaign—with thousands of people marching to demand armed officers leave the city, some physically intervening in arrests, and community members volunteering to patrol their neighborhoods to warn the public when agents are nearby.
But the alliance between Big Tech and the Trump administration on Tuesday interfered with efforts by more than 80,000 Chicagoland residents to show solidarity with immigrants and people of color, as Facebook suspended a community group where people have been tipping off their neighbors when they see US Immigration and Customs Enforcement (ICE) and other federal agents in public areas.
Days after far-right activist and conspiracy theorist Laura Loomer, who has gained considerable influence in the White House despite holding no formal government position, spoke out against a group called ICE Sighting-Chicagoland, Facebook parent company Meta suspended the group to stop its 84,000 members from sharing information about impending ICE raids and enforcement actions.
Loomer wrote on the social media platform X on Sunday that "Big Tech executives" such as Meta CEO Mark Zuckerberg should "use this as an opportunity to be in compliance and to support President [Donald] Trump’s immigration policies, but they aren’t."
She said the presence of the community group was evidence of Zuckerberg's "leftist subversion of Trump and his policies."
Two days later, US Attorney General Pam Bondi announced that action had been taken to stop Chicago area residents from discussing the deployment of federal agents.
"Today, following outreach from the [Department of Justice], Facebook removed a large group page that was being used to dox and target ICE agents in Chicago,” said Bondi Tuesday.
"If the Facebook posts happen to bother Trump, will they still be uncensored, and will their 'free expression' be protected? If you understand what a Trump suck-up Zuckerberg is these days, you can probably take a wild guess."
Bondi repeated a claim by the Department of Homeland Security that immigration agents have faced escalating violence from protesters in Chicago. Few specific examples have backed up the claim, while ICE agents and other officers have been filmed tear-gassing a residential neighborhood; shooting pepper spray at a priest at a demonstration; slamming a congressional candidate on the ground; and holding a journalist on the ground before shoving her in an unmarked car, ramming into another vehicle while speeding away, and eventually releasing her without charges.
The Chicago Sun-Times noted two examples of immigration officers being injured on the job in Chicago recently: one who said his injuries he sustained during a traffic stop that proved fatal for an immigrant named Silverio Villegas González were "nothing major," and another who "hurt his leg chasing a protester."
The administrator of ICE Sighting-Chicagoland posted a screenshot of messages they had received from Meta, which accused the group of failing to follow Facebook's community standards. The group had never been reported or flagged previously.
Meta spokesperson Francis Brennan—a former campaign adviser for Trump during the 2020 election—told the Sun-Times the group had violated Facebook's “Coordinating Harm and Promoting Crime” policy, which bars groups and users from “outing the undercover status of law enforcement, military, or security personnel if the content contains the agent’s name, their face or badge, and any of the following: The agent’s law enforcement organization, the agent’s law enforcement operation, [or] explicit mentions of their undercover status.”
Facebook's policy was revised in 2023; it had previously banned people from sharing explicit identifying information about undercover agents, not mentions of the agencies they work for.
Zuckerberg said earlier this year that content moderation on Facebook had "gone too far" and apologized to Republican lawmakers for previously stopping users from spreading misinformation about Covid-19.
"If the Facebook posts happen to bother Trump, will they still be uncensored, and will their 'free expression' be protected? If you understand what a Trump suck-up Zuckerberg is these days, you can probably take a wild guess," wrote Joe Kukura at SFist on Tuesday.
Zuckerberg was one of several tech billionaires who attended Trump's inauguration in January. Last month he and other Silicon Valley executives attended a White House dinner where they "lavished praise" on the president as they discussed their investments in artificial intelligence and their hopes for a "pro-business, pro-innovation" approach to the technology from the administration.
At the AV Club on Wednesday, Mary Kate Carr said the removal of the ICE Sighting group was "yet another installment of 'How are tech billionaires carrying water for Donald Trump today?'"
"This is the same company that had policies allowing for chatbots to have 'sensual' conversations with kids, but discussing local law enforcement among neighbors is a bridge too far, huh?" wrote Carr.
Keep ReadingShow Less
Most Popular