Joe Masoodi is senior policy analyst at the Dais, a public policy and leadership institute at Toronto Metropolitan University. Cory Searcy is a professor of industrial engineering and the vice-provost and dean of graduate studies at TMU. Patrick Neumann is a design scientist and ergonomist who runs the Human Factors Engineering Lab in the Department of Mechanical and Industrial Engineering at TMU.
For those in the business of selling surveillance tech to employers, the pandemic was a boon. Marginalized and vulnerable workers are the most affected – and with artificial intelligence, surveillance will get only more sophisticated.
From long-haul truck drivers having their eye movements and locations tracked, to fast-food workers being monitored on how quickly orders are served and white-collar office workers having their chats, keystrokes and other communications monitored, the gaze of the electronic eye is both widening and deepening.
Despite evidence about the negative effects of excessive workplace surveillance on workers’ rights – including on privacy, autonomy and dignity – there is relatively little data on this phenomenon in Canadian workplaces, especially since the pandemic. To shed light on the issue, we and our colleagues at TMU conducted a national survey of remote workers, focusing on their experiences with electronic surveillance.
What we found was a disturbing trend, made all the more so when considering the rise of AI and its role in workplace surveillance. If workers are already being surveilled to this degree now, what of the future?
The findings of our survey reveal that seven in 10 employees (70 per cent) have some aspect of their work electronically monitored. E-mail (33 per cent), websites (24 per cent), chats or messages (23 per cent) and phone calls (20 per cent) were the most common aspects of work electronically monitored.
More intrusive methods of surveillance were also reported. About 32 per cent of employees indicated experiencing at least one of the following: location tracking, webcam/video recording, keyboard/keystroke monitoring, computer screen capture or biometrics (facial features, voice and iris scans).
These respondents reported significantly lower levels of job satisfaction and trust in their employer and higher levels of stress. They were also more likely to have a disability, be paid hourly or be lower income.
Such findings corroborate studies in other jurisdictions, including the U.S. and Europe, highlighting the negative psychosocial consequences of excessive surveillance. Moreover, they provide empirical evidence revealing how surveillance disproportionately affects individuals from vulnerable and marginalized communities.
For instance, while one-third of workers (33 per cent) say the levels of electronic monitoring they experience since the pandemic has increased, the rates are significantly higher among workers with household incomes less than $50,000 (41 per cent report increases), those with disabilities (41 per cent) and racialized workers (36 per cent).
The unintended consequences and costs of new surveillance technologies, however, run far deeper. Many different physical and psychological health problems are linked to the loss of control employees often experience over their work with the introduction of surveillance technologies including depression, myalgia, heart disease and various morbidities, not to mention loss of job satisfaction, job strain, increased turnover and absenteeism. AI-enabled surveillance systems will make this worse.
There are growing concerns about workplace surveillance practices becoming overly reliant on copious amounts of personal and sensitive information. A wide array of sensors are increasingly used to carry out such practices as collecting personal physiologic data (cardiac signals, skin conductance, hormone levels, fatigue sensors, voice temperament analysis) that gets fed into AI systems.
Some efforts to address concerns over remote monitoring are already under way. Ontario’s recent legislation requiring large-scale employers to have a written policy on electronic monitoring is a notable example. Our survey, however, reveals that only 30 per cent of these employers have done this.
Clearly, more work is needed. Many workers still lack adequate protections from workplace surveillance practices. Further protections are needed to ensure the least intrusive methods are used. These protections are urgently needed as such practices are increasingly augmented by rapid growth in AI.
The costs in failing to define regulatory guardrails run deep with everyone ultimately paying the price: the company, the worker and society.