Serving the machine? AI in the economy

I was a little girl when the film Terminator 2 was released and it left a deep impression on me. One fine day I started watching the movie on my big brother's video cassette, in secret of course. For the first time, I became aware of one of the favorite themes of the doomsayers, namely the uncontrolled power of machines, and this resulted in many sleepless nights. Even Arnold Schwarzenegger, in all his heroism, was not able to deprogram the threats out of my mind, but fortunately they evaporated over time.

I have not watched the movie since then but it has popped into my mind when AI has become the number one news topic: the biggest threat is that AI becomes self-aware and thus starts controlling mankind. This macro review does not paint the future quite so grimly but it looks at the possible economic effects of AI.

Great hype now surrounds AI (e.g. ChatGPT), which can be seen, e.g., in the Google searches it produces and mentions during the earnings season. The fact that it is becoming more popular has of course been widely recognized but the surprise has been that it has become so common so quickly. In the development of AI we are not talking about years or decades but more about months.

When looking at the economic impact, consensus has been reached primarily on the fact that, in one way or another, AI improves productivity, that fundamental pillar of long-term economic growth. This is also supported by research data, since according to a study by MIT, AI (ChatGPT) helped achieve 40% productivity growth in writing and a clear improvement in quality. Especially in low-productivity jobs, the improvement rate was relatively higher. Thus, in certain sectors and in manual writing, AI can be of great help and employees can focus on more challenging tasks.

It is still challenging to assess the wider impact of the examination, as development is in its early stages and the full potential and areas of application are difficult to identify. A threat frequently mentioned in connection with AI, like other new technologies, is massive job losses and wage cuts. It is difficult to find evidence of this historically: the classic example comes from the banking world, as the expansion of ATMs was once believed to take the jobs of bank clerks. However, this did not happen, but clerks quickly found themselves in other jobs and mainly the nature of the work changed.

However, AI may revolutionize the labor market in the sense that if current routine machine tasks can be outsourced to it, the employee can develop in their job “to a higher level” either by specializing or training and thus increase the number of innovations. Another option, especially in the service industry, is to focus on jobs where a machine cannot replace people (e.g. the care sector). It is likely that the development in the labor market is not of either-or nature, and entire occupations are not immediately lost, but mainly parts of a job can be outsourced to AI.

So what about productivity? Since AI is still in its infancy and has its limitations, it is difficult to talk about the actual productivity leap and a new era of economic growth. However, there is hope that AI could, through the productivity benefits, partly contribute to helping the world economy avoid the dreaded stagflation, where economic growth is slow, but deglobalization, aging of the population, geopolitical risks, and climate change increase inflationary pressures in the economy. When AI is used to organize work tasks more efficiently, at least the labor market shortage can be tackled.

Competition for AI supremacy has intensified at a global level and measured by the volume of investments it is led by the United States and China, and they are well ahead of other countries. As the technological potential is huge, investment will certainly continue to grow strongly. The downside is that AI may become the next trade-policy weapon between the superpowers, in one way or another.