Header Ads Widget

Why the Future AI Will Not Make the World Without Work: By IA Hashmi


AI from computers and robots is again causing waves of technological unemployment fears.

      For a concept to be macroeconomic in the global economy, technology, especially AI, has to be deflationist. It may decrease costs and prices and increase real incomes and demand – this will create new jobs where those destroyed by automation are gone.

     Thus, more and more the key public claim pertaining to technological unemployment has been consistently debunked, whereby macroeconomic new job creation counteracts microeconomic job losses.

      AI is going to improve productivity and generate wealth and new jobs; attention should be paid to sector-specific disruption rather than labor market apocalypse.

In 1983, Wassily Leontief an economist, who was awarded the Nobel Prize for his work on input-output analysis, came to the conclusion that the same fate awaited human labor as the horse after the automobile appeared on the scene – “first reduced then displaced.”

Today, a new wave of gloomy trends concerns the phenomenon of ‘technological unemployment’ Already AI which was expected to bring business changes has stepped into the center of economic discussions. Critics suggest that we’re now approaching a ‘post-work’ world where robots make everything and AI completes all services. Useless humans who are outcompeted by machines are absent from this story of global macroeconomics.

As we noted in Shocks, Crises, and False Alarms the claims that technology leads to total job destruction are not new and thus have an equally poor record of prediction. Concerns with regards worker redundancy is always cyclic to the introduction of newest generation of machines. This threat was once considered so profound that Microsoft’s co-founder, Bill Gates floated an idea to slap a ‘robot tax’ on companies that use robots rather than human beings.

However, in this age of endless technological advancement over the past 80 years where the labor market has been reconstructed – and reconstructed again – the American economy has created 120 million jobs. By 2024, when real wages are rising and the jobless rate stays near historic lows, few people recall Gates’ robot tax notion.

The Hallmark of Technology And Why It Is Said To Be Deflationary?

While still many people accept this story today, not all economists have forgotten the failed visions of technological unemployment. Less dire predictions claim that AI will complement workers – or even lower-paid ones – instead of replacing them. Some counterclaim that it is not easy to replace workers because jobs are bundles of activities, and while AI may perform some activities within that job, it will not be perfect.

Arguments of this kind are valid and are grounded in the microeconomics of work. But macroeconomics is even more convincing in arguing against mass unemployment in the presence of AI.

What is usually left unsaid is that technology is always deflationary in its core essence. When this is widely based, technology reduces costs and prices, promotes increases in real income to consumers and demand for new products – and employment. However, a matter of logic more than luck such rejuvenation of the labor market has taken place from time to time.

It’s perhaps in food that the deflationary effects of technology have been best demonstrated. Thus at the end of the nineteenth century, 40 percent of Americans worked on the farm whereas more than 40 percent of their disposable income was spent on food. Over the next 150 years, ongoing and sequential waves of innovation have resulted in something like 1 percent of the American population toiling on farms. The share going to food budgets on the other hand has declined to about 12 percent of income.



Long-run data of the employment in farm and non-farm sectors of US and percentage of budget spent on food. Image: BCG

Holding costs down gives consumers real income increases: consumers exhibit higher levels of spending on goods and services, older ones often giving way to newer ones, thus creating new employment. Yes, it has always entailed the dismissal of jobs in sectors, and hence microeconomic soreness but it has correspondingly provided a sure way to new jobs and macroeconomic bliss.

Such a track record would not make one a fan of dystopian narratives of technology-induced unemployment; nonetheless. The hit high-tech that often produces dramatic changes over a short time span is easier to identify than new jobs which are constantly emerging in small numbers and over longer periods.

 When it is broadly efficiency enhancing, technology lowers costs and prices, thus increasing people’s real income and demand for new goods and services, or new employment.

 — Philipp Carlsson-Szlezak, Global Chief Economist, Boston Consulting Group & Paul Swartz, Executive Director, Senior Economist, Boston Consulting Group

Will AI defy the trend?

In fact, if we take a closer look at the optimistic projections of AI across all stages of the familiar roadmap that began with new technology and new employment, there appear to be no grounds to expect that AI will put a stop to the process of labor market renewal and adaptation.

                   Cost reduction. It is expected that AI will be able to eliminate labor, more so in service sectors where digitization has otherwise been unable to achieve so. Perhaps the speed and the magnitude of ‘impact’ may not be as huge as what people envision today, but slow and steady erosion of cost (and growth of productivity) is a safer bet.

    Falling prices. If therefore one cannot monopolize on labour-saving technology, the high velocity of cost competition and deflation persists. That is rather far from being unique to food production and has captured various manufactured goods and now staring at the services’ domain. To this effect therefore, policymakers may need to pay keen attention and ensure that market structures are competitive enough so as to cause AI to offered as a deflationary technology.

              New demand. It is often a rise in real income which creates the demand for former rackety goods and services with which nobody was associated when a new technological wave started (would anybody have thought of future social media marketers several decades ago). This demand effect would be an issue of disappointment if consumers pull back and preserve their income increases. But, the subsequent savings accumulation, let alone the apparent savings glut, looks implausible and betrays the historical facts.

             New employment. What if a machine, or an algorithm, can always meet new demand, as the kind of demand illustrated in the kind of variations seen with the price of crude oil? That their comparative advantage disappears and they lose all their politico-economic leverage also appears highly improbable. We can say that AI has labor-enhancing characteristics as credible as the labor-replacing ones. Yet, even if Leontief’s words come true to mean that humans are to go “the way of the horse,” the result would not measure up to a macroeconomic dystopia. Quite the contrary, the hyper-eradicate destabilization of pockets of hyper-inflation would pave the way for revival unequaled in world history.

           The effects overall of AI, and especially of machine learning and viral deep learning, rise to our consciousness level primarily as microeconomic effects that distort and amplify the fluctuations that will ensue. That is not the same as the macroeconomic undertaking that AI represents. Thus, it is absolutely ahistorical and unrealistic to talk about mass technological unemployment. Of course, there are some significant dangers related to AI, but availableness ubiquity careerlessness should not be at the apex of our list of fears.

              What one will most likely witness is a steady rise in productivity and wealth over time with consequent microeconomic crises as is typical of every economic transition. Therefore, let me conclude with a reiteration of what people did not predict; the world will not be without work but it will work differently. 

Post a Comment

0 Comments