May 20, 2026
The AI Backlash Arrives: What Meta, Google DeepMind, and Amazon Employees Tell Us About Trust at Work
Meta’s aggressive push to convert its people investment into AI infrastructure is prompting visible revolt. Last week, flyers appeared in meeting rooms, restrooms, and on vending machines encouraging staff to sign a petition. “Don’t want to work at the Employee Data Extraction Factory?” the pamphlets asked, according to photos seen by Reuters. The petition has now surpassed 1,000 signatures, following months of vocalised discontent on the social media giant’s internal platforms.
This backlash is the result of a series of compounding decisions by Meta – all pointing to a far-reaching agentic AI overhaul that will leave many employees behind. The company cut annual raises by 5% in February 2026, on top of a 10% reduction the year before, while simultaneously raising its AI capital expenditure guidance to $125–145 billion.
Last month, Meta announced it would install tracking software on US employees’ devices to help train its AI model. On 20 May, the company begins notifying affected employees of a 10% workforce reduction, with further cuts expected later this year. An internal memo has also outlined plans to move 7,000 roles into new AI initiatives as part of a broader restructure.
The change is relentless – and it is not coming from a company in distress. This is the strategy of a tech giant going all-in on AI, with little apparent regard for the human cost. Meta is making its mandate clear, and employees are responding with visible protest.
Beyond Meta
This is not isolated to Meta, nor to the United States. Workers at UK-based Google DeepMind have also pushed back against deploying AI without limits. Their bid to unionise follows a controversial deal between Google and the Pentagon that would allow Gemini AI models to be used in the US military for “any lawful purpose.” The deal sparked internal protests, with over 600 Google employees signing an open letter in opposition, and criticism voiced publicly in the press and on social media. Last week, UK-based Google DeepMind employees voted 98% in favour of unionisation to oppose the company’s involvement in military and defence AI contracts.
According to the Union of Tech and Allied Workers (UTAW), the UK unionisation push is part of a global campaign among DeepMind employees that may involve in-person protests and “research strikes” – a form of industrial action involving abstaining from work to improve core AI products.
Elsewhere, Oracle workers affected by the latest wave of US layoffs petitioned for a better severance package; the company refused to move, leaving some employees set to lose hundreds of thousands in unvested stock. At Amazon, where pressure is being placed on staff to use the company’s agentic AI platform MeshClaw, some are reportedly gaming the system. Known as ‘tokenmaxxing’, this behaviour involves automating unnecessary tasks simply to inflate AI usage metrics.
The American Federation of Labor and Congress of Industrial Organizations (AFL-CIO) hosted a Workers First AI summit in recent weeks where leaders from across the labour movement demanded the protection of civil rights. AFL-CIO President Liz Shuler said: “No, you can’t surveil us in the bathroom. No, you can’t steal our data without our consent. No, you are not going to discriminate against us, fire us by app just because a machine told you to.”
Taken together, these movements point to an early but significant shift – where employee discontent is moving from quiet rumblings to labour activism.
The Cost of Disengagement
The scale and spread of these movements has not gone unnoticed by those who study the relationship between employee trust and organisational performance.
Josh Bersin, founder of the Josh Bersin Company, one of the world’s leading HR research and advisory firms, sees the current climate as historically significant. “Employee engagement and trust is at an all-time low – even worse than during the pandemic,” he says. “This is caused by inflation, AI automation fears, and lots of layoffs in the press. AI leaders promoting job displacement have impacted workers, and most employees don’t see value in it yet. Also, the rather ‘inhuman’ way of laying people off – in mass emails – hasn’t helped.”
Bersin is clear that the power shift is real, but that organisations are misreading their own position. “Companies over-hired after the pandemic and now that money is flooding into AI investments, many have stopped hiring or cut back. Given that we want more productivity out of the people we have, EX is actually more important than ever – but workers don’t want any more perks. They want inflation-adjusted pay, flexibility, and some assurance of a long-term career.”
Corporate Cause and Effect
Culture and leadership expert Deborah Hartung argues that what we are witnessing should surprise no one. “Newton’s Third Law applies to organisations just as much as it applies to physics,” she says. “For every action, there is an equal and opposite reaction. You surveil your people while preparing to lay them off, you erode their ethics commitments, you make them train their own replacements – and of course you’re going to get flyers in the bathroom stalls and a 98% union vote. That’s not a coincidence. That’s corporate cause and effect.”
The repercussions, Hartung argues, extend beyond reputational damage. Financial implications are already visible in the AI investment numbers. “Despite $40 billion in enterprise AI investment, 95% of organisations have seen zero measurable impact on profits, and 89% of executives report no effect on labour productivity.”
She adds: “Gallup’s own data shows only 12% of employees in AI-implementing organisations strongly agree that AI has actually transformed how work gets done. Companies are haemorrhaging money on AI tools that their disengaged, distrustful workforce simply isn’t adopting.”
Beyond Big Tech: Why This Matters to Every EX Leader
The organisations making headlines are tech giants. But the dynamics driving employee pushback – surveillance anxiety, AI-driven restructuring, and eroding trust in leadership – are not confined to Silicon Valley or the Square Mile.
Workers in physical industries are already securing concrete protections. The International Longshoremen’s Association won a contract provision prohibiting all fully automated technology. The Las Vegas Culinary union secured an agreement requiring employers to bargain before implementing AI in the workplace. These are frontline workers, not software engineers – and in some respects, they are ahead of many white-collar workforces in translating discontent into formal protections.
For EX leaders watching from the sidelines, Bersin offers a grounding reminder of what the research actually shows. “Trust is always a performance accelerator for employees. Companies that treat employees like replaceable parts usually suffer when they have a business slowdown, because employees won’t lean in. This is why we always advise companies to think of employees as their most important stakeholders.”
The basics, he says, remain unchanged: well-designed jobs that are safe and flexible, supportive management, growth opportunities, fair pay and rewards, and a sense of mission and trust.
