The first thing many young workers now learn about their employer is that it is watching them. Keystroke counters, screenshot intervals, idle-time alerts, badge data: the apparatus arrives before the first performance review, often before the first real conversation with a manager. Whatever the handbook says about trust and empowerment, the software says the operating assumption out loud: we believe you will steal time unless observed.
This is a generational-contract issue more than a privacy issue, because of who bears it. Monitoring concentrates at the bottom: entry-level, hourly, junior remote roles, the exact positions where careers begin and where workers form their durable beliefs about what employment is. A 45-year-old director experiences monitoring as an insult. A 23-year-old in a first job experiences it as a definition.
What the evidence says it does
The American Psychological Association's 2023 Work in America survey found that 51 percent of workers know their employer monitors them, and the monitored are measurably worse off: 56 percent reported typically feeling tense or stressed during the workday, against roughly 40 percent of the unmonitored, and more of them rated their mental health as fair or poor (32 percent versus 24 percent).
The sharper finding is what monitoring does to behavior. Research by Chase Thiel and colleagues, published in Harvard Business Review in 2022, found across two studies that monitored employees were substantially more likely to break rules: cheating, taking equipment, deliberately working slowly. The proposed mechanism is moral disengagement: surveillance transfers the felt responsibility for conduct from the person to the system watching them. Take away someone's agency and you also take away their ownership of what they do with it. The tool purchased to prevent misconduct manufactures the psychology that produces it.
So the entry-level bargain now runs: we will watch you, which will stress you, which we will address with a wellness app, while the watching itself makes the misconduct we feared more likely. It is hard to design a worse trade on purpose.
The alternative was tested decades ago
None of this was unforeseeable, because the inverse experiment has been run. As Ricardo Semler has told it, after his collapse at 25 he began dismantling Semco's control machinery, including the time clock, and eventually stopped counting hours altogether, measuring instead whether the pump was delivered. Absenteeism fell. The company had been surveilling inputs as a proxy for output; when it switched to measuring output directly, the proxy, and the distrust it encoded, turned out to be dead weight.
The same logic ran through the compulsory-vacation rule he describes: thirty days, away, laptop locked, because adults negotiating to work more was not loyalty but fear, and fear, in his phrase, is the cheapest management tool ever invented. Monitoring software is that tool's industrial upgrade: fear, automated, at scale, with dashboards.
What surveillance costs that dashboards don't show
Defenders of monitoring argue that it merely makes visible what managers always informally observed. But informal observation came bundled with context: a supervisor watching a struggling new hire could also see the broken process, the missing training, the bad ticket queue. The software sees only the idle minutes and files them as suspicion. Judgment without context is not management; it is enforcement.
And the cost lands on precisely what entry-level employment is supposed to build. Early careers run on discretionary learning: asking questions that reveal ignorance, trying things slowly before doing them fast, wandering into adjacent problems. Every one of those behaviors reads as inefficiency to a monitoring dashboard. A generation trained under the dashboard learns the opposite curriculum: look busy, hide struggle, never explore. Organizations then diagnose this cohort with disengagement and buy more software to measure it.
The trust question in the generational contract is not answered by values statements. It is answered by defaults. A workplace that opens the relationship with surveillance has stated its prior: this person is a cost to be policed. Workers respond to priors rationally, by giving exactly what is measured and nothing that isn't, and by leaving when someone offers a contract whose opening assumption is better. The companies that will hold the next generation are not the ones with the best dashboards but the ones that made the older discovery: measure the pump, not the person, and the person shows up.
Sources
- American Psychological Association (2023). 2023 Work in America Survey: AI, monitoring technology, and psychological well-being. 51% of workers report being monitored; 56% of monitored workers typically feel tense or stressed at work vs ~40% of unmonitored; 32% vs 24% rate their mental health fair/poor. https://www.apa.org/pubs/reports/work-in-america/2023-work-america-ai-monitoring
- Thiel, Bonner, Bush, Welsh & Garud (2022). Monitoring Employees Makes Them More Likely to Break Rules. Harvard Business Review. Across two studies, monitored employees were substantially more likely to break rules; mechanism identified as moral disengagement / reduced felt responsibility. https://hbr.org/2022/06/monitoring-employees-makes-them-more-likely-to-break-rules