Meta is logging employee keystrokes, mouse movements, and periodic screen snapshots to train AI agents, the company confirmed to staff in internal memos first reported by Reuters this week. The program is called the Model Capability Initiative.

It runs on a list of work-related apps and websites and captures screen content for context. Meta’s stated purpose is to improve AI performance on specific tasks where models still struggle, such as navigating dropdown menus and using keyboard shortcuts.

Also listen on Apple Podcasts

Meanwhile, Meta is preparing to cut as much as 20 percent of its workforce, with the first layoffs reportedly set to begin in May. Workers are being recorded doing their jobs in order to teach systems the work those same workers currently perform.

The company has committed up to $135 billion in capital expenditure for 2026, most of it directed toward AI. The workers generating the training data are not receiving additional compensation for it. There is no reported opt-out mechanism.

Every office worker reading this should ask one question: is my employer doing this too?

The scale of what employers are planning clarifies why that question matters. A working paper published by the National Bureau of Economic Research, drawing on a survey of nearly 750 corporate executives conducted with Duke University and the Federal Reserve Banks of Atlanta and Richmond, projects AI-driven job cuts in 2026 at approximately 502,000 roles. The 2025 figure was roughly 55,000. The projected 2026 number is nine times higher.

Get the eBook of the Substack series that dissects the narratives Tech uses to build its AI Empire. $9.95 flat fee for PDF or ePUB formats; no subscription required. 3.5-hr reading time.

The people building artificial intelligence did not invent their ideas. They inherited them.

The researchers also document what they call a productivity paradox: executives perceive AI’s productivity gains as larger than the data actually shows. Companies are cutting based on expectations that have not yet materialized in their revenue figures.

That gap between expectation and reality shapes the bind workers currently face. A survey of 2,400 knowledge workers across the U.S., the U.K., and Europe — conducted by enterprise AI firm Writer and research firm Workplace Intelligence — found that 29 percent of employees admit to actively sabotaging their company’s AI rollout. Among Gen Z workers, that figure rises to 44 percent.

The most common reasons cited are fear of job loss and frustration with tools that do not work as promised. The sabotage takes forms ranging from refusing to use approved tools to entering proprietary information into public AI platforms without authorization.

However, workers who resist AI adoption face termination risk: 60 percent of executives surveyed said they are considering cutting employees who refuse to adopt AI tools.

Future Forwarded is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

Workers who embrace the tools and demonstrate their capabilities face a different risk — they may be accelerating their own replacement. Seventy-seven percent of executives said workers who remain AI-resistant will be excluded from promotions and leadership consideration.

The person building the technology making the most widespread use of it has his own assessment of where this leads. Anthropic CEO Dario Amodei warned at Davos in January 2026 that AI’s reach across cognitive work is categorically different from earlier automation waves.

Factory workers displaced by industrial automation could, in theory, move into service or office roles. If AI is moving into office roles simultaneously, Amodei argued, the adjacent sector that absorbed earlier displacement does not exist. “The technology is not replacing a single job,” he said, “but acting as a general labor substitute for humans.”

So, one of the people most responsible for building the technology is the one making the case that the safety valve is gone.

Get the eBook compilation of Substack articles examining how the AI invasion already happened. You just weren’t invited. $9.95 flat fee for PDF or ePUB; no subscription required. 2-hr reading time.

Comments

Get full access to Future Forwarded at futureforwarded.substack.com/subscribe

Get full access to The AI/Labor Report at ailabor.substack.com/subscribe