AI ‘Kills the Radio Stars’. The Cyclical History of Work

Talking seriously about AI and work is hard, because it is hard to talk seriously about work. Reference to a generic ‘place’ of work, typical of both economists and labor lawyers, is of little use if we want to talk about how jobs are changing in the face of changing technology. As management and organizational scholars try—too timidly—to say from Taylor onward, jobs are either specific or cannot be the subject of either design or management or training. This is because specific jobs are complex sets of specific skills each of which reacts differently to technological change.

As a consequence, much of the debate is faux-ideological (battles between technophobic neo-Luddites versus technoenthusiasts only pretend to be furious). And it results in seasonal complaints and Cassandra-style predictions that are as generic in substance as they are specific in numbers (‘x.y million jobs will disappear’). Another, even sadder part of the recent debate, is then an attempted sales pitch that starts by waxing lyrical about the technology of the day (AI today, PC before, coal before that) in order to sell us a ‘package’ that we most likely do not need.

In any case, nothing to help us responsibly and positively manage the actual transformation of jobs. Which is big. Identitarian. Unsettling. For businesses. Humans. And the community.

But let’s put nihilism aside and let’s try to speak the truth.

First and foremost, AI is an engine of transformation. As on dozens of other occasions in history, it is a technological engine that rattles us. As Kubrick’s bone told us in vain. AI is the new coal, most interesting for coal experts. But for us, labor experts, it is ‘just’ a new bone. A new coal. We, labor design enthusiasts, must still make the effort to disassemble labor into its activities and competencies and begin redesigning it from there. More or less alarmist predictions about jobs ‘disappearing’ can only be futile. And, in fact, beyond the absolute numbers, if you then look at the list of jobs you always find the usual Minzbergian categories of clerical jobs that will no longer be there. Let it be said with the utmost respect for some prestigious institutions that churn out the lists of these disappearing jobs, but the first three are the same jobs that could be expected to disappear when the first calculator arrived. 

As more food for thought, AI is an engine of change, and nothing is more disruptive to humans than change. In recent research it is among the top reasons for worker stress (along with the ‘green’ transformation). And this is no change on the past either. In the twilight zone, monsters are born. And shedding light is our duty. With data. And here is, meanwhile, data at the macro level that is important to recall: between 1960 and 2017, the number of workers in the United States nearly tripled. And a specific analysis on PC showed that since 1980 it has led to the net creation of nearly 16 million jobs. But then above all, we need to collect data in firms. Collect real data about real jobs and show the specific professional families how many and what new skills need to be learned. Clearly the implications on specific jobs, or rather specific skills, are variable. And some skills become obsolete or completely useless and others grow and get stronger. It is important to start with the specific lists.

Third and last and most important reflection I want to say here now. Businesses and communities can and should prepare action plans for skills transformation. But there is only one possible action plan and it is called mass training (or retraining) system. More precisely, mass sector-specific training plans. Scalable. Involving millions of people. We need to start from company value chains and the impacts of AI on different specific activities. And from those activities retrace back jobs. And disassemble those jobs into specific skills focusing training on those skills. Train people at the workplace, with rigorous plans for structured on-the-job training.

Indeed, the central issue of technological transformation is, as ever, its impact on specific skills. It is the skills needed to do the jobs that change. It is the skills that ‘disappear’ in the sense of no longer being needed because they are replaced by the new technology of the day. And it is the new skills that need to be the subject of mass training plans. For these new skills we do not need to activate ill-advised pots of windfall money. But we must activate detailed implementation plans, mixed public and private plans for retraining the unemployed and the employed. And with coordinated effort between the public system and businesses, we must adjust the education plans of the youngest, starting with vocational schools.

This essay may look very prosaic in a debate that nowadays is used to warming up to more adventurous words. But today is not the time for adventure. It is the time to be responsible. And serious.