In a recent conversation about the affect of artificial intelligence on employment on Professionals Australia’s new discussion board, a commenter made a point about the amount of training required to obtain and hold down a job now compared to the amount required in the past. Once upon a time (the nineteenth century, say), people could be employed by age fourteen and continue to work in this profession for the remainder of their lives. Now, we don’t expect anyone to be fully trained for employment until the age of twenty-one or so, and we’re endlessly warned of the need for re-training throughout our working lives in order to stay employed. The commenter observes that schemes such as the universal basic income are supposed to help with this (by providing an income while one takes time out to re-train) though he himself isn’t sure that a universal basic income is the best solution.
I’ve noted before on this blog that we haven’t really grappled with how all of this training is to either paid for or made time for. Many commentators have a vague idea that people will need more training in order to perform tasks that haven’t been automated, and will need to continue re-training as even more tasks get automated over time. Yet conversations about robots taking jobs and/or technology creating new jobs rarely tackle how this training is to be done and paid for, or how it might affect the way we work.
For a start, training has a significant cost, not just in terms of building schools and employing teachers, but also in the amount of time that people spend doing something that doesn’t immediately produce any goods or services. We’ve gotten use to paying these expenses up until workers graduate from university or apprenticeships, but I’m not sure we’ve gotten used to meeting meeting these costs throughout life let alone conceived of a plan to actually make it happen. In Australia, at least, we certainly don’t have a formal scheme on anything like the scale of the school and higher education systems (though I know that Singapore has SkillsFuture).
Of course, from an economic perspective, time spent in training is also a kind “investment”: one hopes that investing in workers’ skills will pay off with greater productivity in the future, such that there is a net gain in productivity even considering the “unproductive” time spent in training. We can and do pay for universal schooling and subsidise post-school education precisely because the modern workers that graduate from such institutions can produce so much more than fourteen-year-old factory hands using nineteenth-century technology. But we don’t much think this way once we’ve left formal school.
Then there’s the question of what sort of training any given worker out to undertake. Once upon a time people imagined robots taking over physical labour, leaving emotional and intellectual labour for humans, which they were frequently presumed to prefer. But more recently it has become apparent that robots can perform many sorts of routinised work, be it physical or intellectual, such that they are wonderful at spreadsheet calculations but terrible at folding clothes. So should I study liberal arts, say, in the hope that people like Jeffrey Sachs and Anant Agarwal are gearing up to hire philosophers, or should I brush up on how to clean a toilet?