AI workplace surveillance grows as firms track employee activity
NEW YORK, UNITED STATES — A new wave of workplace surveillance is sweeping United States companies, with employers now tracking keystrokes, mouse movements and digital activity not just to measure productivity but to harvest the data needed to train AI agents that may eventually replace the workers being watched.
According to a report from Business Insider, Meta is among the latest tech giants to deploy an internal tool that monitors employee activity to train AI systems, joining JPMorgan, which is tracking software engineers’ AI use through dashboards.
While AT&T, which is using technology to monitor office attendance — a shift in the U.S. Government Accountability Office flagged in a September report on the rapid growth of employer surveillance.
Why employers are moving beyond productivity tracking
The strategic logic centers on what companies call “digital exhaust” — the trail of clicks, messages and decisions workers leave behind as they do their jobs.
That data is now seen as a competitive asset, because real-world examples of how employees solve problems are among the most valuable inputs for building effective AI agents.
Meta said the data collection is designed to help AI systems understand how workers complete tasks, with safeguards in place to protect sensitive content. Some Meta employees have raised concerns about how closely the company will monitor their work, but the practice is expected to spread as more companies face investor pressure to show returns on heavy AI spending.
“This is the evolution of workplace surveillance — from measuring work to learning how to replace it,” said Dan Schawbel, managing partner at research firm Workplace Intelligence.
The trust gap reshaping the U.S. workplace
The data collection trend is colliding with growing concerns about transparency. Emily Rose McRae, senior director analyst at Gartner, credited Meta for disclosing its monitoring program to workers, noting that too often employees aren’t aware of what companies are collecting — and that learning about it after the fact “feels like a betrayal.”
She also cautioned that even rich activity data may not capture the full picture of how knowledge work actually gets done.
Schawbel said the broader pattern reflects a “trust erosion” between workers and employers, with many staff accepting surveillance because they have little leverage in the current job market.
Some, he added, see cooperation as a way to extend their own runway in roles that may eventually be automated.
Schawbel said workers can now be a company’s “best technology asset” as AI training pulls institutional knowledge directly from frontline staff.
For the future of work, the implications run deep. As U.S. employers race to build AI agents on top of their own workforce data, the line between employee productivity and AI training data is dissolving — a structural shift that will reshape how performance, privacy and job security are defined across white-collar roles.
The outsourcing industry stands to absorb meaningful share in this transition, with business process outsourcing (BPO) providers offering trained annotation teams, AI quality assurance specialists and compliance reviewers becoming essential partners for companies that need clean, ethically sourced training data without the legal and reputational risks of mining their own employees.
