Valued at work

A growing number of organisations are using new digital technologies to track their workers. They collect information about them and then use algorithms and artificial intelligence to analyse the data so that they can gauge workers’ suitability, assess their health and appraise their work. We call these technologies digital monitoring tools. Organisations expect these tools to help them make informed decisions and thereby optimise the value of the work.

In this report, commissioned by the Dutch House of Representatives’ Social Affairs and Employment Committee, we explore what new methods of tracking, analysing and giving feedback to workers mean for job quality. We look at three key dimensions of this concept (earnings quality, labour market security and quality of the working environment) and at broader social and ethical aspects. We investigate a series of questions: What does tracking mean for worker privacy and for creating an inclusive workplace? Do these tools in fact increase the value of the work? Our study consists of desk research, a literature review, interviews and an analysis of relevant legal frameworks. The study was carried out between June 2019 and March 2020, and contributes to the knowledge base under the agenda theme ‘Changing labour market’ of the Committee.

The specific impact of digital monitoring technology for job quality depends on many different factors. For one thing, the digital tools are not all alike. This report shows that there is now a vast array of tools available. We have grouped them into three categories: tools for staff planning and hiring, tools for managing and instructing workers, and tools for their support and development. Organisations can choose the level at which they wish to use the tools (individual, team or organisation) and at which of these levels they will provide feedback. They can also analyse data for retrospective purposes (historical trends in remuneration or absenteeism), to look for connections (for example between leave and absenteeism), or to predict trends (for example, future absenteeism).

Because organisations differ, each tool can impact job quality in a different way. That impact also depends on how organisations choose to deploy the technology. Despite the differences between organisations and tools, our research has revealed some broad trends.

Digital monitoring tools can be detrimental to workers

We see that various digital tools can be detrimental to workers. This may be especially the case when they influence decisions on a candidate’s or employee’s eligibility for a job, promotion or contract renewal. There are also concerns about privacy and discrimination. Specifically, as such tools can collect sensitive data, including email, location, movement and sleep patterns, facial expressions, and even hereditary traits. As a result, they may impinge on worker privacy. And although artificial intelligence and algorithms have the potential to mitigate discrimination in job application procedures, the risk of discrimination often remains. 

Little proof of validity in some cases

In addition, the tools have only limited success when it comes to worker analysis. They are often used to analyse complex matters, such as suitability, motivation, health and productivity and base some of these analyses on questionable connections, for example between facial expression and personality, or between DNA and competencies. There is scant evidence for the validity of various tools. Our respondents caution organisations to beware of ‘cowboys’ operating in this market.

Human beings are difficult to ‘capture’ in data

Organisations must realise that it is difficult to gauge the value of work even with the help of digital tools, algorithms and artificial intelligence. For example, research has yet to reveal the extent to which we can predict human behaviour, such as the likelihood of an employee leaving an organisation or of a candidate being suitable for a job. Assessments using artificial intelligence rate applicants based on the attributes of the most successful employees, but in doing so, they create a similar-to-me-bias. Moreover, it may be underestimated how much those working quietly behind the scenes contribute to someone else’s high performance.

There is a risk that the analysis will focus only on what is quantifiable and ignore other essential activities or personality traits. Although a numerical workplace analysis can be useful, it may overlook other valuable aspects of work. Moreover, the underlying view that humans can be ‘captured’ in data produces too limited a notion of what valuable work is.

Monitoring technology changes work processes and workplace relations

Using digital monitoring tools can change work processes and workplace relations. Choosing to track a certain aspect of work puts that aspect under the microscope. The organisation and its workers then act accordingly, potentially producing unintended and, occasionally, counterproductive effects. While it may seem productive to monitor call centre performance by tracking the number of calls employees handle in a given time frame, the result may well be stressed-out employees and dissatisfied customers.

Organisations that focuses too much on maximising worker efficiency may end up driving up the workload and restricting the professional autonomy of their workers. Are employees at liberty to assess what a customer needs, or to consult with their colleagues when necessary? Giving employees a tool that counsels them on how to cope with their workload gradually shifts the responsibility for addressing such pressure from the organisation to the individual. This means that digital monitoring can be detrimental to cooperation, shared responsibility and even job satisfaction.

Broad public and political discussion about data in the workplace

The quest for a ‘data-driven’ workplace exposes a dominant rationale in which organisations use data to understand people. We are moving towards a labour market in which quantitative data will be crucial for predicting behaviour and for decisions influencing employment opportunities. The responsible use of digital monitoring tools therefore requires critical reflection, clear communication and dialogue. Organisations should not only be concerned about protecting privacy, preventing discrimination and addressing workload, but must also realise that their digital tools impact workplace tasks, processes and relations, and that monitoring affects the value of work. Despite the growing availability of such tools, the ‘data management’ of organisations is not always adequate and hinders the use of this type of technology. This means that relevant stakeholders still have the opportunity to anticipate these developments.

The expanded possibilities to digitally measure, analyse and give feedback to employees can have adverse consequences for employees and for the quality of work in general. It affects, among other things, the opportunity to participate in the labour market and how work is assessed and rewarded. It also increases how much organisations know about employees. Moreover, the instruments change employment relationships. Not only between organisations, employees and technology providers, but also between employees themselves. 

Moreover, the underlying view that people can be captured in data provides too limited a picture of what valuable work is. Where organisations try to optimise the value of work, there is a risk that the value of work will actually decline.

This calls for a broad social and political discussion about the desirable use of monitoring technology and the value of work.

The Rathenau Instituut would like to see a broad public and political discussion about the appropriate use of monitoring technology and valuable work. To encourage this discussion, we have identified three starting points.

1. Discuss the appropriate use of data analysis in the workplace

The Rathenau Instituut invites employers’ associations and trade unions, online platforms, workers and technology vendors to discuss the use of data in the workplace on the following basis:

• Be realistic about the opportunities and limitations of technology and have an open discussion about what constitutes valuable work. Prevent potential job impoverishment.
• Be cautious about using special categories of personal data and automated or semi-automated decision-making processes, including in job application procedures. Clarify how algorithms work, and apply the principles of proportionality (is the breach of privacy in proportion to the intended purpose?), and subsidiarity (is there a privacy-friendly alternative available?).
• Be aware of the unintentional or undesirable effects of the changing relationships between workers and organisations and between colleagues, suppliers, customers and regulatory bodies. Pay close attention to the position of technology vendors. Explore the capabilities and terms and conditions of trusted third parties who can manage worker data.

2. Set quality requirements for digital monitoring tools

There is still too little hard evidence on the validity of various tools. This is specifically the case when it comes to recruitment and selection, precisely an area where decisions based on these tools may be detrimental to workers and job applicants. A new code of conduct for recruitment and selection procedures was introduced (by the NVP, the Dutch Network for HR professionals) in February 2020 that requires algorithms to be validated and transparent and potential risks and shortcomings to be clear. There are no further rules governing how this is to be accomplished and neither is it clear who is monitoring compliance and how. The relevant stakeholders should flesh out the details as quickly as possible.

3. Invest in active supervision and enforcement

The current legal frameworks set requirements for the use of digital monitoring tools but it is unclear how certain legal principles should be applied in the workplace. That is why regulatory bodies such as the Dutch Data Protection Authority (DPA) and the Inspectorate SZW (the Dutch labour inspectorate) play an important role in enforcing these frameworks.

Four issues merit special attention in that context:

• Actively enforce the law when it comes to tools that do not meet legal requirements and offer more detailed explanations when standards prove ambiguous. Work with other regulatory bodies in doing so.
• Be especially attentive when it comes to processing special categories of personal data. The DPA and the Inspectorate SZW could jointly clarify which data (including health data) are required to comply with the Working Conditions Act, good employment practices and the General Data Protection Regulation (GDPR).
• Devote extra attention to addressing workload. Among other things, consider that digital monitoring technology can lead to a gruelling work pace.

• Keep a close eye on the fairness of selection procedures, even when using artificial intelligence.