Global judgements and ideas.
In preparing the text for my current European Parliament STOA project Workplace Monitoring & Surveillance in the Digital Era, I have been looking at my 2018 book Quantified Self in Precarity: Work, Technology and What Counts pp. 146 – 150, where I talk about where and how EPM is used today, from people analytics to sociometrics to facial recognition. The research for my book was all done before the GDPR was introduced…
Chapter: ‘Unseen Labour and All-of-Life Surveillance’ pp. 146 – 150
Phoebe V Moore
Electronic performance monitoring (EPM) typically involves email monitoring, phone tapping, tracking computer content and usage times, video monitoring and GPS tracking. The data produced from new technologies can be used as productivity indicators; indication of employees’ location; email usage; website browsing; printer use; telephone use; even tone of voice and physical movement during conversation (see wearable technologies section). Perhaps the longest history of EPM is seen in call centre work (Taylor et al., 2002) where various types of surveillance facilities lean on working practices. Emotion tracking is standard activity in Indian centres (Van Jaarsveld and Poster, 2013) and practices are increasingly normalised internationally.
In the early 1990s, a US Senator on the Labour and Human Resources Committee, in hearings on the Privacy for Consumers and Workers Bill (S. 516) warned that ‘unrestrained surveillance of workers has turned many offices into electronic sweatshops… electronic monitoring should not be abused… Employees should not be forced to give up their freedom, dignity or sacrifice their health when they go to work’ (Collins, 1991). Nonetheless, by 2010, an estimated 75 percent of American companies were shown to monitor employee communications and other at-work activities (Ball, 2010). The estimated change in the US market use of technology to monitor employees rose by 43 percent between 2007 and 2010 (Harpers Index, 2010 quoted in Schumacher, 2011: 138) and we are now in a period of what seems to be ‘limitless worker surveillance’ (Ajunwa et al., 2016).
Our facial images have been captured and stored by institutions for decades, from the library to the motor vehicle department. We are accustomed to some forms of identify cards and the terrible photo that only the immigration officer in an airport should ever see. Now, a machine collects facial images during every trip through the airport. While finger prints require more explicit consent and have normally only been collected if the hand belongs to someone under suspicion, our biometrics are now seemingly up for grabs. CCTV is nearly ubiquitous in some countries. Faces are becoming the new ‘bar codes’ for all kinds of things (Introna and Wood, 2004). Face + + Cognitive Services is one such facial recognition system being rolled out in China now. Face + + can detect and locate faces in images; compare faces; search and locate similar faces in a large collection; identify key points of face components it calls ‘landmarks’; and analyse attributes including gender, emotion, head poses, and eye status (Face + +, 2017). In China, the platform has been used by the ride-sharing company Didi. The Face + + customer testimonial section on its website states that
Didi is the largest ride-sharing company providing transportation services for over 300 million users across China, and the parent company of Uber China. With facial recognition service by Face++, Didi succeeded in building a safe and easy riding experience for both riders and drivers. Face Comparing offered great help in solving the problems of fraud and cheating (Ibid.). A simple look has already replaced a fingerprint to unlock phones. Iris detection to unlock doors seems less and less like a far-fetched scene from a science fiction film.
In the workplace, likely areas for EPM with such functions are in work design decisions, such as the use of Sociometric Solutions for placing furniture in offices to facilitate optimal movement around offices. This ‘solution’ tracks movements around offices and people’s tones of voice. Taking it to the next level, video cameras could de-anonymise such trackers. Virtual reality training could be an area for leadership and management training. VR is already used by the U.S. Army for training soldiers. The Training Industry Magazine reported in 2017 that VR management courses are available for managers in training to master effective skills using VR, where simulations for face-to-face coaching session with a problem worker help a budding leader develop responses to employees by playing back recorded sessions, which are also scored and ranked compared to colleagues and participants (Young et al., 2017). Social media can be used to ‘predict personality type’ from the Big Five traits of ‘extraversion, agreeableness, openness, conscientiousness and neuroticism’, and coaching sessions can help leaders to recognise their own personality types to improve use of social media. Information about employees’ personality types can also be used to feed into coaching sessions and help people understand why people ‘engage in specific behaviours that are revealed through other means’ (Ibid). This report indicates that social media analytics are not perfect, and there are ethical questions about acquiring data used to determine personalities without users’ consent. The other possible trend is wearable devices that can record physiological data like heart rate which, the authors state, can now give an ‘objective assessment of our stress state’ (Ibid).
EPM can be used to micromanage employees and invade privacy, lower job satisfaction, increase stress and leads to low-trust, negative work relationships (Schumacher, 2011). Human dignity is at stake in the context of rising workplace surveillance and, as Rothstein pointed out, in relation to workplace monitoring:
At work, human dignity is denied by treating the employee as a mere factor of production with fixed capacities and vulnerabilities determining her behaviour and ignoring both the worker’s individuality in the face of statistical probabilities and the human potential to overcome or compensate for physical obstacles. The worker’s dignity is denied when she is treated as a mechanism transparent to the view of others at a distance and therefore manipulable or disposable without the ability to confront the observer. (Rothstein, 2000: 383–84; also see Gantt, 1995)
Direct productivity monitoring enabled by software installed in work and personal computers reveals a new EPM technique. Examples include RescueTime, Toggl, ATracker and My Minutes which have been introduced in real estate design workplaces such as in the case of the Quantified Workplace study (Moore, 2018). The constant onslaught of communications and information and expectations to personally manage work that was once done by another specialist in the company in timetabling and accountancy has led to 24/7 working lives (Bogost, 2013). The ‘overwhelmed employee’ checks her mobile devices up to 150 times a day and suffers from information overload, inability to find time to reflect and even just to think, leading to employee disengagement and undermining productivity (Hodson et al., 2014). People can now never really switch off, even while asleep or on holiday. These pressures are compounded in a new world of work where employees are ‘always on’, or even ‘hyper-employed’.
The power of tracking work with EPM becomes even more attractive to employers with the possibility for the aggregation of ‘big data’. Productivity can be monitored with increasing accuracy, offering detailed second-by-second frequency with the use of ‘people analytics’ that extrapolates from the data made available. People analytics is made possible by the reduced costs of data and information processing. It is of interest to business because it is seen to reduce costs in service-provider selection and/or workplace re-organization and restructuring. It is the dominant selection method in the ‘sharing economy’ (discussed below). Implicit to the use of these technologies and the large amounts of data produced, as well as analysis from people analytics, is that the type of activity and the length of time spent on activities can be inherently linked to a qualitative judgment about a worker’s performance; information that could be used in appraisals or hiring and firing decisions. Intimate performance dashboards provided by most EPM technologies incorporate contextual information obtained from tracking devices such as levels of physical activity, level of stress, or presence and absence scores. The data itself is seen as the indicator of value. However, as pointed out by Angrave et al. (2016: 7):
…the process of modelling and creating dashboards and traffic lights is not value neutral but depends on dominant paradigms and perspectives within accounting and operations management, which themselves reflect ideology, politics and power.
These practices are rapidly superseding other forms of management methods as data produced is seen to be a reliable indicator of productivity. New EPM is very different from traditional methods (Jeske and Santuzzi, 2015). Reliance on metrics from tracking devices potentially dehumanizes employees who are reduced to a collection of activity timestamps. The associated measures can result in biased performance evaluations; pressures for increased work or work intensification; reduction of autonomy (linked to privacy concerns) (Bhave, 2014; Haque, 2015); and perceived intensified control over individuals’ work (Jeske and Santuzzi, 2015). These pressures lead to reduced commitment and lowered job satisfaction. EPM has the very real potential for uses of psychological bullying, and employees are concerned about the possibilities for workplace control based on what can be considered new surveillance methods (Ball and Margulis, 2011; Rosenblat et al., 2014).
In 2014, 36 percent of workers in a Gallup poll (Harter et al., 2014) said they checked work email outside of normal working hours and this number is not likely to drop. Employers are legally permitted to monitor workers’ emails and internet browser history, with few restrictions. Facebook’s terms of service are more binding than most governments’, stating that ‘you will not share your password… let anyone else access your account, or do anything that might jeopardize the security of your account’. There is little actual legal protection over these violations of privacy and employers have been increasingly known to ask potential employees for social media log-in information (Beesley, 2016). ‘Workplace spying’ (The Week, 2015) can, unsurprisingly, lead to anxiety and psycho-social discomfort, or what could even be called ‘violence’ (Akhtar and Moore, 2016). Research has shown that when employees perceive they have privacy; it actually improves productivity: a feature of the ‘transparency paradox’ (Bernstein, 2012).
There are difficulties in adapting the concept of respect for private life into workplaces. The right to privacy is taken to cover professional or business activities and employees’ communications, thus implying that employers may not, in principle, interfere in these areas. However, as Kilkelly (2001) states, such interference is acceptable in certain circumstances – notably ‘in the interests of national security, public safety or the economic well-being of the country, for the prevention of disorder or crime, for the protection of health or morals’. This raises questions about the extent to which workers may be monitored and the relationship between their rights and employers’ prerogative. All of these questions are linked to rapid automation of work and a prescient question of machine management. How far will it go?