by Josiah Ober and John Tasioulas
'The use of AI at work needs to be responsive to the intrinsic value of work and workplace democracy'
In the realm of work, we are free to reject Aristotle’s belief that working at the direction of others, or for the sake of others’ enjoyment (e.g. in musical performance: Politics 1341b9-14), is inherently illiberal, and as such degrades our reasoning capacity. But we can nonetheless acknowledge that much work has, historically, been degrading both physically and psychically – dull, repetitive, unimaginative, exhausting. AI can help us to eliminate degrading labour by transforming work into an expression of our prosocial capacities, our human excellences. Still, we contend that work is a profoundly important context for human fulfilment.
Work, human flourishing and AI
Engaging in work activities can have instrumental value, for example, by generating valuable goods and services and an income for the worker, by honing skills that are useful beyond the workplace, such as in personal life and political activity, etc. But it can also be a source of great non-instrumental value that is constitutive of human flourishing. Key among these values, we believe, is that of achievement, which consists in the valuable exercise of our powers in meeting difficult challenges for a worthwhile end. Meanwhile, collaborating with one’s work colleagues and customers towards the realisation of a common project is also an important form of friendship. A politically significant by-product of achievement is a justified sense of self-esteem, a quality that can play a vital role in sustaining a democratic ethos in which citizens feel able to look each other in the eye as equals, contributing through work to the common good of society as a whole. This depends on a general recognition of the valuable role that different forms of work, whether of hand, brain, or heart, play in sustaining the common good.
If work is a vital domain for human flourishing in the modern world, the question arises of how we should respond as AI systems increasingly acquire the capacity to perform work activities. Of course, there are many forms of work – notably, dangerous or demeaning or otherwise distasteful tasks – that we should be happy, and may even be obligated, to delegate to AI systems. But there has never previously been a technology that has the potential to replace human work activities on such a significant scale. This goes well-beyond ‘routine’ or ‘mechanical tasks’ to include white-collar occupations, such as journalism and legal services. A 2017 Oxford-based study concluded that 47% of all occupations in the United States are capable of being “computerised” in the next 10-20 years.[1] One needs, however, to take such dramatic claims with a grain of salt. As Oren Cass has noted, one of the jobs the authors of this study supposed could be fully automated is that of school bus driver:
From a tall enough ivory tower, or a heady corner of Silicon Valley, the claim about school bus drivers might seem to make sense. What could be easier than driving a school bus? The route is the same every day, it’s short, and it gets canceled for snow. For parents, though, the idea of locking twenty kinds in a self-driving vehicle for half an hour, with no adult supervision, sounds dubious at best. [2]
More circumspectly, a McKinsey study found that, although just under 5% of occupations are fully automatable, around 30% of all work tasks in 60% of occupations could be automated.[3] So the potential threat posed by AI to jobs seems significant. Nor can we confidently assume that, as with past technological innovations, new jobs will emerge to replace those eliminated by AI.
Human flourishing in a ’post-work’ world
Among tech leaders, such as Mark Zuckerberg and Elon Musk, the idea of a Universal Basic Income has found favour as a solution to the challenges of a post-work world. [4] Now, there is reason to doubt that the replacement of human workers with AI systems will lead to productivity gains large enough to generate a UBI that covers lost income – a great deal of automation represents what Acemoglu and Johnson have called “so-so innovation”, displacing workers without any significant rise in productivity or the quality of goods and services. Much here, of course, turns on the contested issue of whether Artificial General Intelligence is a feasible, desirable, and reasonably imminent development.
But even if a UBI could address the economic inequalities threatened by technological unemployment by redistributing pieces of a larger economic pie, it would not itself remedy the massive reduction in opportunities for achievement, as well as the consequent loss of a source of friendship and democratic self-esteem. And, from a purely practical standpoint, it seems unlikely that the UBI scheme would be sustainable in the long run in a world in which a democratic ethos has been frayed by the status inequalities and the attrition of skills consequent upon the displacement of large numbers of citizens from productive activities.
One obvious thought here is that the loss of opportunities for achievement will be compensated for by the greater leisure time people will have to pursue other values in a world without work, values such as friendship, the pleasures of artistic appreciation, travel or fine dining, religious observance, the playing of games, etc. John Maynard Keynes greeted the prospect of a jobless future in this vein in his 1930 essay entitled ‘Economic Possibilities for our Grandchildren – ‘we have been trained too long to strive and not to enjoy’. [5] But even this cautious optimism seems overly optimistic to us, elevating passive enjoyment to a status it cannot have in an Aristotelian framework that prioritises the cultivation of the excellences of character and intellect in its account of human flourishing.
Another alluring thought advanced recently is that the lost opportunities for achievement afforded by work can be replaced by achievement in the playing of games. [6] Some have taken this idea even further, suggesting that we will be occupied in playing even more exciting and challenging games in virtual reality, involving such things as the virtual capacity to fly unaided, and that a life spent doing so can be just as good as a life spent in the ‘real world’. [7] We have here, it seems, the elements of a philosophical prescription for the good life in the metaverse.
But this play-based virtual utopia rests on deeply contestable theses. First, that the primary value of game-playing is that of achievement, rather than play itself (when factory workers play an impromptu game of football in their lunch-break, is the primary value realistically secured thereby that of achievement in athletic prowess, rather than the fun of a kick-about?). Second, that the value of achievement is to be understood in an explicitly anti-Aristotelian way, as focussed entirely on the process involved (skill in overcoming difficulties) rather than also on any substantive good achieved by that process, since typically the objectives of games (e.g. putting a ball in a hole, crossing an arbitrary line before others) are trivial or valueless. Only when the value of ends comes into view – the valuable goods and services produced through work – can we grasp why the sense of achievement derived through work as a nurse, plumber, teacher, farmer etc. cannot for the great majority of people be satisfactorily replaced by proficiency at activities such as chess, golf or table-tennis.[8] And it is precisely the disconnect from suitably valuable ends that explains why many fail to find fulfilment in otherwise well-paid, but pointless, white collar ‘bullshit jobs’, however challenging the difficult the tasks they involve. [9]
Finally, the thesis about virtual reality brings us back to even more fundamental themes broached earlier, about our essential nature as human beings with a specific biological form. This is adapted to living in the natural world along with other human and non-human beings with their own biological form. Work is not just as a source of achievement, friendship and self-esteem, but also as a way of engaging with and understanding a reality that exists independently of us rather than an artificial reality synthesised by human beings.
As David Wiggins has written, elaborating a neo-Aristotelian account of the meaning of work:
Acts or activities that apply what he [Aristotle] calls a rational principle aim at something worthwhile by drawing upon faculties and dispositions whose exercise gives pleasure (a distinctive, associated pleasure) to the doer and enlarges also – here I reach beyond Aristotle – the doer’s understanding of the realities we inhabit. That is to say that the exercise of these faculties or dispositions affords both practical understanding of those realities and the satisfactions that we attain by learning to wrestle or struggle with them. [10]
The idea here is that work affords a distinct form of understanding that emerges through contact with an independent and potentially recalcitrant physical reality: in work, we engage with that reality in a way that deploys our rational faculties to create goods and services that satisfy human interests. This is related to the idea found in Hegel’s dialectic of master and slave, which in turn influenced Marx, that work is vitally significant for human self-realisation because it involves the struggle to transform nature into a humanised domain of culture (Bildung).[11] By humanising work in this way we more fully realise our own human potential, which includes enabling us to conceive of ourselves as moral agents possessing equal rights.
Worker participation in shaping the deployment of AI
In short, how AI technology should be integrated as tools into the work environment to enable both the flourishing of workers and the common good of society remains a serious challenge. The simple idea that we should promote the use of AI technology, including by systematically displacing human workers, to stimulate economic growth and then redistribute some of the wealth gained to the jobless through a UBI, turns out to be simplistic. A focus on economic growth, through such measures as GDP, fails to reflect all the value we derive from work, notably, achievement, friendship, self-esteem, and an understanding of the world around us by engaging with it to produce goods and services.
Going back to Aristotle’s scepticism about a life of work due to its supposed lack of self-direction, an important element in addressing this challenge will be giving workers a greater voice in determining the shape of their workplace. This includes the role that AI technologies should play at work (consider, here, the German co-determination system, whereby workers’ councils have a say on a company’s supervisory board). One of those beneficial roles for AI technology should be that of enhancing worker participation in corporate governance, as opposed to it being part of a system of surveillance and control geared to extracting the maximum economic value from ‘human resources’. AI ought to be used to enable workers to contribute to work processes and products using their heads as well as their hands - inverting Henry Ford’s (perhaps apocryphal) regret that the hands he hired in his automotive factories came attached to heads. AI tools should enable workers to be participants in the organisation and management of work environments, in much the same ways that it could enable more participatory forms of political democracy.
[1]CB Frey and MA Osborne, ‘The Future of Employment: How Susceptible Are Jobs to Automation?’, Technological Forecasting and Social Change no.114 (2017): 254-80.
[2]Oren Cass, The Once and Future Worker: A Vision for the Renewal of Work in America (Encounter Books, 2018), p,69.
[3]James Manyika and Kevin Sneader, “AI, Automation, and the Future of Work: Ten Things to Solve For,” McKinsey Global Institute Executive Briefing, June 1, 2018, https: //www.mckinsey.com/featured-insights/future-of-work/ai-automation-and-the-future-of-work-ten-things-to-solve-for.
[4]Catherine Clifford, ‘Elon Musk: Robots Will be Able to Do Everything Better Than Us’, CNBC, July 17, 2017 https://www.cnbc.com/2017/07/17/elon-musk-robots-will-be-able-to-do-everything-better-than-us.html
[5]John Maynard Keynes, ‘Economic Possibilities for our Grandchildren’, in Essays in Persuasion (Harcourt Brace, 1932), pp. 358-373
[6] See John Danaher, in Automation and Utopia: Human Flourishing in a World without Work (Harvard University Press, 2019), p.236, referencing the work of Thomas Hurka, ‘Games and the Good’. Proceedings of the Aristotelian Society supp. Vol.80 (2006), pp.217-35.
[7]For the philosopher David Chalmers, “virtual realities have comparable value to nonvirtual realities”, hence life in a computer simulation can be just as meaningful as life in a non-virtual world. D. Chalmers, Reality+ Virtual Worlds and the Problems of Philosophy (Allen Lane, 2022), p.328
[8] For these two objections, see John Tasioulas, ‘Games and the Good’, Proceedings of the Aristotelian Society supp. Vol. 80 (2006) 237. In line with the first objection, however, play is a substantive good itself that the playing of games can realise. See also John Tasioulas, ‘Work and Play in the Shadow of AI’, in D. Edmonds (ed), AI Morality (OUP, forthcoming).
[9]David Graeber, Bullshit Jobs: A Theory (Allen Lane, 2018).
[10]David Wiggins, ‘Work, its moral meaning and import’, Philosophy 89 (2014), p.479
[11]G.W.F. Hegel, The Phenomenology of Spirit (OUP, 1976), pp.111-119.