The critical flaw at the heart of all visions of the “future of work” in which artificial intelligence makes most professions obsolete and drives billions out of work may be summed up in the phrase, “internal anachronism”.
This concept is best illustrated by a major defect of most science fiction movies.
Take Luc Besson’s Valerian, for instance. A civilisation that has learnt to travel across hyperdimensional space still uses staccato-firing weapons and relies on natural zoological species to replicate physical objects.
In the Wachowski Brothers’ Jupiter Ascending, a civilisation that has conquered the light barrier still uses wolverine DNA to promote aggression in its soldiery, whilst, wait for it, growing feathery wings on their backs for air mobility.
The problem stems from the difficulty of true multidisciplinary thinking. To project well into the future, one needs to understand a vast array of disciplines, scientific and humanities-based, and deeply grasp how findings in one field impact and grow atop developments in other fields ensuring some degree of harmony in technological advancement.
That is how come cooking in microwave ovens and talking to other people using wireless devices appear like starkly divergent cultural realities and yet provide a common defining hallmark of the late twentieth century. In fact, only a span of five years (1967 to 1972) separated the commercialisation of both technologies to the point where household ubiquity was only a matter of time.
The principle at play here is that a civilisation that has mastered electromagnetic radiation would apply it to remarkably diverse aspects of their culture, triggering powerful trends in multiple areas of research and application. Indeed, “microwave sterilisation” in the food sciences (which takes preservation of food to a whole new level) seem very removed from the nodes used in the Internet of Things, until we start seeing wireless sensors in microwaves, as is the case with the Tovala smart oven. Smart ovens, like Tovala, should eventually enable anyone to translate virtually any recipe into an adequate meal within a very short period of time.
The interplay of wireless technologies in this manner thus mean that at a certain point in the near future, the impact of wireless on our nutritional lives will not be limited to food-ordering apps. It is very likely that in a world where sensor technology has advanced to a point where it is truly ubiquitous, food ordering apps would be obsolete and “smart cooking” would become a better reflection of the “proper harmony across the state of relevant technologies”. If that is the case, then showing a food ordering app being used in a city completely swathed in driverless car infrastructure would constitute a case of “internal anachronism”.
I deliberately chose a very subtle example to illustrate the nuances. A much easier example would be an intelligent ambulance driving a person to an emergency theater rather than implementing the stabilisation and recovery techniques in situ.
If the long preamble has served its purpose of explaining the essence of the idea, then it should be easy to see how it applies to the dystopic visions of the future of work.
Simply put, the kind of general, human level, intelligence expected of computers in the next couple of years can only happen if there is an acceleration of a vast multitude of fields, both because AI at such an advanced state should dramatically boost research in every other field and also because to solve some of the complex problems confronting AI today would require advancement in a wide range of adjacent and not-so-adjacent fields, from networks to smart materials to micropower engineering.
The impact on our economy would be transformational to the point of opening up whole new frontiers in subsea habitation, space outposts, urban greening, species restoration, subterranean complexes etc etc. It is not simply that “new industries” will create new types of jobs. It is also that existing industries will grapple with scale issues of unimaginable proportions that cannot be automated a priori until the social, economic and technological interests have aligned enough to allow automation.
And to the extent that many of these new opportunities would first require human investment decisions, political clearance, and economic reconfiguration to take off, the employment cycle would follow the pattern of high human uptake followed by a productivity plateau and thereafter automation to improve returns on investment. Though the cycles would grow shorter, the new waves of “employment expansion” will also come faster and more intensely, even as the growth in human populations slow (a universal phenomenon of industrialisation).
A distorted understanding of how inter-disciplinary cross-fertilisation powers the growth of technological capacity, and how economic and cultural cycles interweave with waves of structural change, greatly underplays the power of “new frontier” exploration to open up new opportunities for humans to engage productively with the construction of new societies, even when the forces of change seem uncontrollable.
If you think that this is a far-off vision, then you get the point. Human-replacement level AI is a far-off vision too, because its emergence presupposes the vision I have described above. The same forces that can undermine, frustrate or drag that “massive scale” vision are the very ones that shall slow down the impact of human-replacement level AI and encourage instead many intermediate steps of human-complementary AI.