Even the economists don't understand the AI revolution

This week, Prime Minister Jonas Gahr Støre distributed the “KI billion” to six national centres based in universities. Over five years, the six KICs will receive NOK 1.2 billion to stimulate more knowledge production about KI, both technically and socially.

The KI billion has already worked. It has generated enormous interest from the research communities and stimulated subjects to collaborate in ways they have not done before. I'm in one of them. TRUST, at the University of Oslo, spans the sciences, social sciences, law and humanities and has dozens of partners in the public and private sectors.

But when I look at who has received crackdowns, unfortunately there is one professional group that has, to a small extent, asserted itself, and that is the social economists. It is striking all the time that KI is expected to have enormous consequences for the economy and society at large. Social economics is, alongside law, the dominant management discipline and KI is of great interest to a growing number of social economists internationally.

Economists' lack of interest in ICT directly affects the authorities' knowledge base. The Fiscal Policy Committee does not mention KI in a word, nor what consequences rapid automation of labour will have on the economy and Norwegian society. IN The Perspective Message Looking at developments up to 2060, marginal gains from automation are projected.

That KI will have marginal gains stands in stark contrast to the estimates of leading experts. According to thousands of experts on KI it is more likely than not that by 2047 we will have KI systems that perform all work tasks better and cheaper than humans.

We already have access to highly capable AI systems, which are on par with the world's foremost experts in coding and mathematics. They possess an extreme amount of knowledge about difficult questions and is already helping to moving the research front. With the deep research capabilities of ChatGPT and Claude, one can, in a matter of minutes, write reports that would take weeks for a PhD student to create.

The development also shows no signs of stopping. If we believe the leading companies, within a few years we will have KI systems that can do almost all office work.

The economic consequences are expected to be large. On the one hand, KI may be the gift package that seemingly solves many of our problems. It can free up labor that can become the hot hands that are in demand in the health sector. But even in an optimistic scenario, the Norwegian economy will face one of its biggest restructuring challenges since the Industrial Revolution.

The absence of good analytics means that the authorities are just as in the fog as they were when ChatGPT was launched and turned the education sector upside down.

That economists miss the most important technology trend of our time is striking all the while the social economy's best explanation for economic growth is precisely technology development. They also ignore a nascent non-fiction on the economic and societal implications of KI.

In the book Power and Progress Last year's Nobel laureate in economics, Daron Acemoglu, argues that KI will weaken the power and economic situation of workers. He believes we should strengthen unions, reduce taxes on work and invest more in labor-supplemental technology to avoid automation becoming an evil.

Other leading economists, such as Anton Korinek and Chad Jones, predicts explosive economic growth if the KI experts are right that the machines will provide a multiplication in the supply of skilled and hardworking knowledge workers and the profits go to the capital owners.

Norway is not prepared for this development. Our existing welfare schemes, which rely on unemployment being a marginal phenomenon, are in no way sized for such a scenario. Nor is the tax system rigged to fund a welfare state should large swathes of payroll income lapse.

If value creation is shifted from labour to capital and many workers struggle to readjust themselves, how should the tax system and the welfare state be aligned? Is it wise, for example, to employ the employer levy, which makes it more expensive to employ people than machines, in a phase of rapid readjustment? Are labour market measures and continuing education programmes adapted to a situation where a large part of the population has to learn to perform new work tasks?

Are the key institutions of the labour market, such as trade unions and the tripartite cooperation, equipped to ensure both increased productivity growth, but also the distribution and balance of power between the workers and the owners of capital?

To be prepared for what is to come, we need our best professionals to give society and politicians the best possible understanding of the changes that may come and the associated research. That we still have few empirical studies to lean on and that the future is uncertain is not a good reason to ignore what may or will come.

Social economists are good at helping us deal with societal problems we've largely already solved, like how to make marginal changes to National Insurance benefits and tax rates to increase the supply of labor.

What society needs most, however, is for our wise heads to spend time on the problems we don't have a solution to, such as how best to deal with the coming KI revolution. Large parts of the knowledge sector are now mobilised. When will the economists get on the field?