Work hard or work smart?

Brian YS Wong May 09, 2024 22:35

I have long grappled with a quandary to which I have – perhaps unsurprisingly – very little solid ground to fall back on in answering. Indeed, this is a quandary that is shared by many of my fellow academics, non-academic peers, and, of course, students out there.

In the age of automation, what are the kinds of jobs that would be replaced the most quickly? This, in turn, gives rise to a more fundamental question – what are the kinds of individuals whom automation cannot replace quickly, if ever?

The intuitive observation is that working hard doesn’t get one very far anymore. Blue-collar work and purportedly low-skilled white-collar work (I believe the latter is a misnomer – “low skills” may be more easily automatable, but there is no reason to think that the skills demanded are inferior, more accessible, or more straightforward in any regard) have been the first to go under the second and third waves of mass industrialisation and mechanisation. One could be the most industrious lad in the yard, but when an automated machine can do one’s job at a pace much higher than and with costs much lower than one, that is when one clearly pales in attractiveness to prospective employers.

Yet this intuitive answer flies in the face of three trends. The first, as pointed out by the economist David Autor, is the growing demand and rising wages amongst blue-collar workers in the US throughout the course of the pandemic and in recent years – which speak to the transformations taking to the American industrial regime and economic structure. The second, is that so-called ‘high-skilled’ white-collar jobs (e.g. accountancy, and even lawyers and some investment bankers) are increasingly displaced by algorithm-based AI models that can do their jobs “for them” – if not in directly replacing and substituting for them, then most certainly through reducing considerably their unique value-added to their employers, who would look to the burgeoning AI industry for models and systems that can meet their needs.

And, most fundamentally, as outlined by various significant consultancy reports and articles speculating over the case for a Universal Basic Income, automation could very well reduce the total number of working hours that workers have to endure in the future – that is, with greater automation comes greater leisure time for those whose labour is now freed up. Working hard is not only no longer a prerequisite for success – it is also losing the implicitly fostered association with being an indicator of success. Just think of all those Wall Street bankers who gloat over their long, intense working hours, and equated such arduous working conditions with their supposed excellence.

Of course, this last point assumes that those with less ‘work time’ will then spend it on ‘leisure’. Yet this dichotomy is bogus – firstly, there’s the tricky, inevitable component of ‘care time’, especially in societies with ageing populations, or booming with young children. Then there’s the question of whether the unemployed and the jobless could in fact enjoy such time – painful, stressful, anxious time spent worrying over how to make ends meet could not be deemed leisure. And finally, there’s the fundamental question of how much leisure we in fact need – why should we assume that work and leisure are strictly dichotomous? Can’t we hold that the two are intertwined and enmeshed with one another, and jettison the macabre, reductionist worldview promulgated by Marx and his ilk on labour and leisure?

In any case, perhaps the answer rests with “working smart”. The hypothesis, or so it goes, is that smart people can weather automation and come out on top. The more intellectual, the more multi-disciplinary, the more creative and capable of generating novel ideas, claims, and works of art one would be. In short, work smart by going creative – turn to the arts, music, and the cultural sphere: there’s always something in it for everyone, no?

Yet this view omits the fact that with Generative AI that can de facto emulate some of the best and most pronounced aspects of human creativity (through diffusion techniques and complex probability maneuvers), AI-powered automation is also coming for the ‘jobs’ of scriptwriters, playwrights, and, indeed, opinion editors. We may still be enamoured of the op-eds of the likes of Tom Friedman, Martin Wolf, or Shoshana Zuboff for now – but that’s because of their enduring brands and name recognition.

Yet what if there comes a day when AI can be ‘smarter’ than us, too? Not just in terms of book-smartness and factual recall, but also when it comes to intellectual profundity – or the ability to mimic the signals that our heuristics recognise as signs of so-called intellectualism? Let’s call it pseudo-smartness. The truth is, we don’t need AI to be smart in order to pose a threat to us; all we need is for it to possess the pseudo-smartness – the ability to talk about topics off the top of their heads, to improvise and to deliver solid-sounding soundbites and claims, and, indeed, to develop and broadcast deepfake TikTok videos that are lapped up by many who would willingly comply amongst our youth.

Working smart won’t get us very far either. What’s needed here, is “Working as a human”. I am reminded of Hannah Arendt’s observations and musings on space travel – that when humanity can finally venture beyond the Solar System, there will come a point in time when the future astronaut looks back at us, and realise how small, how insignificant, and how trivial humanity looks, “just like rats”, from the furthest reachest of the traversable outer space.

And it is at that instant that the astronaut would make up his mind to come home, for humanity cannot stomach the thought of not being the protagonists of this world. And it is precisely this fixation upon the human, the real, and the authentic, that will make or break an industry and occupation going forward. Some humans just can’t be replaced with machines, because their jobs require the operators to exhibit the flaws, the defects, and the imperfections that make human three-dimensional carbon-based creatures, as opposed to silicon-based or virtual entities.

Brian YS Wong

Assistant Professor, HKU