Artificial Intelligence has finally come of age because of machine learning and the availability of scalable computing power on the cloud. But here’s a new problem that AI researchers needs to solve for: public discourse about this potentially world-changing technology is being monopolized by three tribes of experts, each one with their own agenda. The unfortunate outcome of this discourse, as framed by those tribes’ agendas, is the painting of a dystopian future where humans become virtually obsolete.
The first tribe are the technology companies; their seller’s mantra is old and tested: “automate to reduce costs and increase profitability”. This mantra, however, has an implicit corollary: human worker layoffs. This corollary is reinforced by the musings of the second tribe, the economists. By applying dubious neoclassic methodologies to labor market data they predict employment apocalypse within the next ten years. And then a third tribe comes along - a late-comer - and jumps on the bandwagon of AI doom: the politicians. Having sensed the next great opportunity to extend the power of government (i.e. their power), they see no alternative than borrowing more money from future generations in order to fund the grand idea of “Universal Basic Income” (”free money for all”) as the antidote to the supposedly mass unemployment that awaits us.
What all three tribes seem unaware of are two major risks that none of them has accounted for. The first risk is what happens if we do not automate the economy. Although automation due to AI, robotics etc. has the potential to fully or partly automate many jobs, it is important to put this fact into a macroeconomic and microeconomic context. We live in an era of low unemployment, low productivity, cheap money and bad demographics. I would also argue that, despite the buzz and hype around AI, we also live in an era where companies invest very little in technology, and generally find it easier to throw relatively cheap (read "unproductive") labour at tasks than take the risk of capital investment in automation. Automation technology is a necessary but not sufficient condition for widespread automation to happen. Companies need to take the decision to automate. The perception of cheap labor in much of the west and the continued focus on running lean is a serious impediment to that decision being taken any time soon. As always, money talks; and money nowadays is cheap - and labor cheaper. Despite the push for automation very few companies in the West are doing anything about it, preferring business as usual and a wait-and-see “strategy”.
A robot army of noodle makers coming near you soon...
But the West is no longer the uncontested leader on this planet. Judging by China's bullish advance in automating its economy, one can only surmise that western economies will make the leap to serious automation only after their economic competitiveness begins to seriously flounder, due to further loss in productivity as well as because China will be doing almost everything better. But when western businesses wake up it might be too late. By then, western countries will have massively increased their debt in order to finance healthcare and pensions for their increasing older populations, taxing the minority of young people to such an extent as to virtually eliminate the desire to work hard, or succeed. Even today, given the high taxation in most developed countries, it makes little sense to be more productive if what you will end up making is all but taken away from you by the taxman. Unless western economies start automating now, their people will end up as the poor of the 21st century.
There is a second as yet unaccounted for risk too: by projecting automation as something to be fearful of the public in western countries may revolt against AI. It will not be the first time that a promising technology was rejected by the public; just think of GMOs. If the same thing happens with AI, then the economic decline of the West will come much faster, and be a lot worse.
If you are a worker living in a western country, fear not of the robots taking over your job. Fear they will not do so quickly enough..