Just International

AI and the Workplace: Maintaining workers’ dignity

By Dr Jaspal Kaur Sadhu Singh

Work has varied connotations of importance. If what we do is simply an economic means to sustain our lives and our families, it may not define who we are as we may be driven by other meaningful pursuits. However, we may differ from this view if work is seen as a vocation. Regardless of its meaning, it is undeniable that work flourishes our lives in some form and we expect the preservation of dignity and treatment as workers.

The tiresome exploitation of the pandemic has raised the issue of adoption of AI for oppressive and ominous uses at the workplace. I often have to qualify my views on the use of AI-enabled technologies — I am indeed a proponent of its altruistic abilities to advance humanity, but with caution.

Any technology may be the antidote or the poison and AI is no different. AI augments human decision-making by automating certain decisions and tasks that previously required human cognition. Human intervention in prediction and decision-making may not be required where we can have autonomous AI with advanced computing power, enlarged storage capacity through cloud computing, and the collection and processing of large data sets popularly referred to as “big data” to train the algorithm or alternatively, from which the algorithm learns. The most remarkable feature of AI is its ability to learn from these datasets and make decisions.

Countries promote AI innovation to take advantage of AI-driven technologies, particularly automation and augmentation of work processes. Naturally, such an agenda will catapult a nation towards economic competitiveness and drive it to ensure that its workforce possesses the necessary skills that are shaped and assisted by AI-enabled systems. Experts are divided on the impact the use of AI may have on jobs and wages. I am convinced by the view proffered by Paul Daugherty and James Wilson in their book Human + Machine — that automation and augmentation of work processes will allow workers to do their jobs better and faster. For instance, the utility of AI-systems in healthcare during these difficult times has produced remarkable results. AI-systems trained by thousands of images of patients’ scans have relieved pressures on medical professionals by reducing the time taken in diagnosing and treating critical illnesses and saving lives.

However, another view that there are a class of workers who may be displaced by automation when they are unable to reskill themselves exists. This inability is due to the cost and opportunity of reskilling, lack of aptitude, and most importantly, the dereliction by employers to develop training and development strategies to reskill their workforce. If a nation wishes to adopt automation through AI-assisted technologies, national policies and human resources initiatives sophisticated enough for the nation’s workforce to transition must be paramount. A U.S. survey from May 2021 reveals that companies are planning to make significant investments in deploying AI-systems in their organisations, an initiative propelled even faster by the pandemic that has profound implications on the employment relationship.

Denise Rousseau in her 1995 book, Psychological Contracts in Organizations: Understanding Written and Unwritten Agreements, introduces a relationship between individuals and the organisations framed within a psychological contract comprising beliefs shaped by the organisation which are “terms of an exchange”. Psychological contracts embed employees’ perceptions and goal-setting which is mutually productive to both workers and their employers. These contracts may be either transactional or relational – the former being a relationship of a short duration, monetary-focussed, with little mutual enhancement; and the latter, possessing a stronger mutuality element, imbued with socio-emotional and monetary values. Developing a relational contract that is enriching to the employer and employee will result in a higher sense of purpose, mutual trust and confidence, and benefit the socio-economic success of the organisation, national industrial harmony and workers’ dignity at the workplace in the long run.

Does the current acceleration of automation and adoption of AI-systems at the workplace weaken the relational contract between employers and employees in terms of job engagement and trust? Whilst recognising how AI may help workers or act to improve safety of workers, is removal of workers’ autonomy acceptable? Or are we simply to measure its utility through benefits reaped by employers? Or will there be ramifications?

The use of AI-systems at the workplace is evident in people management functions and automation of work processes — we have read of the use of AI-recruitment processes which candidates must pass before receiving an interview. Candidates are required to upload videos of themselves responding to questions on a platform that uses AI-systems to determine a person’s suitability for employ based on facial expression. Proponents of such methods claim that the system is allegedly free from bias and is more accurate in determining a person’s suitability than human decision-making authority. This is however, not the case. In 2018, Amazon’s version showed a bias against female applicants and essentially discriminatory.

What if AI is to be used to make decisions on performance, promotion, and even termination? When AI-enabled cameras were installed in Amazon’s vehicles, the company’s position was that the cameras improved the safety of its drivers and delivery network. All good so far. But what if this footage is then used as a constant monitoring tool as well as a performance metric? Working conditions in Amazon have been considered oppressive for many years and the constant recording of drivers is draconian too, especially when the drivers’ failure to accept the camera installation will result in their termination.
I am frequently surprised when speaking to AI-developers and deployers who dismiss the ramifications of the use of AI and the ethical, moral and legal questions. We need to engage with these questions and be present in the debate as ‘the absent are always in the wrong’.

There is nevertheless some element of delight as we pause at the threshold of opportunity. Upon identifying the discriminatory issues in its recruitment AI, Amazon proceeded to adopt a proposed new test for ensuring fairness in algorithmic modelling and data driven decisions called ‘Conditional Demographic Disparity’ developed by Professor Sandra Wachter and her colleagues at the Oxford Internet Institute.

Elsewhere, workers’ organisations are taking initiatives to manage AI-adoption at the workplace. In the UK, a Trade Union Congress (TUC) report from March 2021 highlighted the implications of AI-adoption on employment rights; of erosion of equality where there is discrimination, compromising privacy and data protection where there is surveillance, as well as the impact of the technology on the physical and mental wellbeing of workers. The TUC report also highlights larger issues concerning the balance of power between employers and workers, and dignity and democracy at work. The TUC’s survey of workers revealed the workers’ lack of awareness of whether AI-enabled systems were being utilised which is indicative of the lack of transparency and democracy. The TUC has produced an awe-inspiring manifesto – Dignity at Work and the AI Revolution – that aims to ensure that adoption of technology at work benefits all parties, helping employers navigate the AI opportunity and promise without disregarding their workers. It presents fundamental values to be respected such as equality, work/home boundaries, transparency and explainability, including proposals for new rights such as the right to human review and human contact when decisions are made about workers by predictive AI systems, to be recognised. The manifesto is a preliminary template for an ethical framework for employers when introducing AI technologies which impact their workers. The consideration of these values and rights ensures there is involvement of human agency, or in machine learning parlance, the human-in-the-loop. The relational contract between employees and their organisation will strengthen, striking a balance between dignity and development.

These are volatile times indeed but it is not a time for being inattentive to values of empathy and humanity. Let us not suffer the folly of being impervious to countervailing evidence that dignity of work is worth preserving.

4 June 2021

Dr Jaspal Kaur is a member of the Executive Committee of the International Movement for a Just World (JUST). She is also a Senior Lecturer at the Faculty of Law & Government, HELP University, Kuala Lumpur. She is a co-founder of a knowledge portal, ai-doctrina.info currently under development, with a quest to raise the ethical and legal AI maturity in the country.