Workers and the rise of the algorithmic boss
The character of labour markets and the future of work differ significantly across countries. Workers’ Day, also known as May Day or Labour Day, commemorates the struggles and gains of workers and is observed in many countries on 1 May. Global geopolitics, economics and climate change developments raise the thorny question of what the foremost struggles are that face South African and global workers today.
The International Labour Organisation views the aim of decent work to ensure that different groups in the labour market have equal opportunities in employment and income, safety and security at the workplace, social protection, rights of association (union membership) and social dialogue. Against this background, the issue of generative artificial intelligence (GenAI) and the workplace or, stated more strikingly, “the rise of the algorithmic boss” is relevant.
There is significant speculation about the effect of GenAI for the future of work and workers — the nub of such effect is sometimes framed as displacement vis-a-viz risk-shifting. Debates on the future of work initially focused on automation, the introduction of new technologies at the workplace and therefore the loss of jobs because of technological innovation.
Another aspect of this innovation is equally important, namely the quality of jobs in future labour markets. The trade union movement has adopted the phrase “Dignity at work amidst an AI revolution” because they predict that if left unchecked, “the use of AI to manage people will also lead to work becoming an increasingly lonely and isolating experience, where the joy of human connection is lost”. People spend a large portion of their lives performing paid work, at the workplace or increasingly from places other than a traditional workplace, and therefore it matters greatly what the quality of (power) relationships, interactions and conditions at work are.
The World Economic Forum predicts that by 2026 more than 80% of organisations will have used GenAI in interfaces or applications running in production environments but that only 16% would have done so successfully. Employers must consider how to use GenAI effectively from an operational, financial and people perspective. Surveys show that almost a quarter of employees expect GenAI to replace their job in the next five years. Despite this anxiety, research shows that in the short- to medium-term, AI won’t replace many jobs; on the contrary, it may create, not diminish, workforce opportunity. In fact, James Bessen from the Boston University School of Law convincingly argues that one cannot simply assume that productivity-improving technology necessarily leads to job losses, and that while productive technology may or may not decrease aggregate employment, it will probably have disparate effects on different industries at different times.
Research and surveys do indicate that, without doubt, AI will lead jobs to be redesigned to include new responsibilities, such as interacting with GenAI tools, and certain jobs and tasks will be disproportionately affected. For example, the 2023 Gartner survey predicts that GenAI will play a role in 70% of text- and data-heavy tasks by 2025.
Some of the emerging concerns regarding GenAI and the quality of jobs that should then be on the agenda of policymakers and regulators include workers’ privacy and electronic monitoring, insidious forms of discrimination (explicit and implicit biases) hidden behind a technologically “objective” facade, practices such as people analytics and the use of big data and AI to manage the workforce (management-by-algorithm), helping workers transition from some industries/occupations and some regions to others, risk-shifting from employers to workers through using AI-driven tools for algorithmic scheduling, task redefinition (also what is regarded as compensable work and time), loss and fraud prediction (often resulting in predictive targeting and preventative measures), and incentivisation of productivity (including invasive forms of monitoring and “inactivity reports”).
Scholars, including Valerio De Stefano and Antonio Aloisi who co-authored Your Boss Is an Algorithm: Artificial Intelligence, Platform Work and Labour, have convincingly argued that in this context a necessary approach for suitable policymaking is “countervailing worker power”. Such approach centres around strategies that increase data transparency, privacy and fairness underpinned by workers’ rights, and empowering worker representatives to develop their own knowledge of GenAI as well as their capacity to effectively negotiate the adoption and impacts thereof. It is evident that workers experience real difficulties to assert their individual rights in the context of algorithmic management systems, so collective labour rights, and especially collective bargaining, are the most effective and proven tools for workers to influence the so-called “distribution of benefits or costs from the AI- and data-driven ‘digital revolutio’”’. To this, one should add information and consultation rights of worker representatives.
Other strategies include imposing joint liability on companies for breaches of labour law associated with algorithmic management technologies that they develop or market, and developing parallel regulation setting minimum standards for decent work for the worker concerns described earlier. What we need is a multi-level response to address the effects of AI on employment relations. This response could include protective mechanisms (minimum standards set through regulations or sectoral collective agreements) and participative standards (giving employees or their representatives bargaining rights and resources to govern processes through direct participation).
In 2019, Jeremias Adams-Prassl of Oxford University wrote that the “real challenge lies in harnessing the unequivocal potential in the trends which will shape tomorrow’s work, whilst ensuring that no one is left behind in enjoying decent and sustainable working conditions”.
The law does not simply respond to technological change but as the Cambridge academics Simon Deakin and Christopher Markou have shown, it also facilitates and mediates it. Each decade has presented workers with challenges and labour law has often been responding to these challenges rather than leading with the introduction of gains for workers and their families. Although the present AI revolution will continue to bring far-reaching consequences to workplaces and workers, it appears that many of the “new” challenges are rooted in historical developments (such as atypical work and platform work, the nature and scope of managerial prerogative and worker rights to fair labour practices). Hard-won advances in law and policies may be sustained and accelerated in many instances through thoughtful and tailored adaptation and implementation.
Solutions will have to build upon existing responses while being more agile in introducing new tailormade responses to GenAI specific challenges, such as management-by-algorithm. Deakin and Markou correctly propose that GenAI has the capacity to undermine existing forms of regulation, while creating the space for new ones. Several jurisdictions, including the European Union, have adopted artificial intelligence Acts with provisions relating to classes of AI systems that are deemed high-impact systems (for example, employment related decisions or court or administrative body decision-making). Key to this classification is the severity and extent of potential adverse impacts, including on human rights and social harms, while it is recognised that overlap between these classes and existing regulatory regimes may mitigate the need for specific inclusions or exclusions.
In South Africa our workforce is diverse, and work is performed across the formal and informal economies. The scope of structural unemployment, especially among the youth, is well recorded. These stumbling blocks may explain why we seemingly have been slow to respond to the challenges posed by the rapid introduction and widespread use of GenAI in the labour market. But this is an instance where the law should play a leading and facilitating role in ensuring decent work through mediating and diffusing the effect of GenAI on both job security and the quality of workers’ jobs and personal rights. The opportunity to imagine what the future of work should look like should therefore be taken up with more urgency and intent.
Professor Nicola Smit is the dean of the Faculty of Law at Stellenbosch University.