Investments and Leadership Are the Key to Unlocking AI’s Potential

New forms of AI will accelerate growth in labor productivity in the coming decades. Europe has fallen behind, which threatens competitiveness and thus the foundation of our welfare, points out Jens Lundsgaard.

There’s a duality in the message from Jens Lundsgaard, Senior Fellow at CIP Foundation. He has just taken up the role of Director for Global Presence at Novo Nordisk Foundation, coming from a position as Deputy Director responsible for research, technology, and innovation at OECD, where he has, among other things, examined AI’s impact on economic growth. He expects it will take time before new forms of AI lift the economy. At the same time, he has a crystal-clear message: Massive investments in AI are urgently needed. Otherwise, Europe will fall even further behind.

“There has been much more massive investment in the USA and China, and given the significance new forms of AI will have, it will be decisive for the economy and prosperity. Much larger investments are needed from the European side. The sooner, the better,” he says.

Generative AI is among the most advanced forms of AI, where the algorithm can be trained. While earlier versions of AI can process existing data, generative AI can create new content.

And the question is not whether generative AI will have a decisive impact on productivity. Rather, it’s whether the impact will be significant, very large, or enormous. So even when, like Jens Lundsgaard and OECD, one is cautious in expectations, there is no doubt that the integration of new AI will be crucial for competitiveness over the next 10–20 years.

OECD’s analysis indicates that various forms of AI over the next ten years can raise labor productivity by nearly 0.5–1 percent annually. This is in addition to the existing productivity improvement, which is around 1–1.5 percent per year. Such figures may not sound like much, but it is significant, as it means that labor productivity in the coming years could grow 50 percent faster than before, says Jens Lundsgaard.

Broad Impact

Other researchers and think tanks expect even greater effects from AI. The International Monetary Fund (IMF) points out that AI provides annual productivity gains of 1–1.5 percent, while McKinsey has previously worked with a broad estimate from 0.5 up to 3.5 percent.

“Across these very different estimates, it’s clear that new forms of AI will have a decisive impact on productivity. When you look at the major players in the global economy, it will be crucial for competitiveness in the coming years to be able to reap the benefits of AI. Unfortunately, the status is that Europe lags far behind the USA and China,” says Jens Lundsgaard.

In many areas, generative AI is still an immature technology, so it’s hard to pinpoint exactly which areas will be affected—but perhaps even harder to point out which areas won’t be affected.

We can already see that AI is having a broad impact, asserts Jens Lundsgaard. Whether it’s development and research in new products, optimization of production, diagnostics in healthcare, or communication and language processing, there are already areas today where companies are working with AI solutions. And we are only at the very early stages; many companies have not yet truly begun to exploit the possibilities.

More Than Party Songs and School Assignments

The tip of the AI iceberg became visible and popular when ChatGPT appeared two or three years ago. Everything from party speeches and translations to school assignments could be solved with the right prompt.

But generative AI goes much further back and has a much broader significance. Especially because AI will have an extreme impact on research and development.

“In many cases, generative AI will significantly reduce the time from the initial idea phase to when a new technology can go into production,” says Jens Lundsgaard.

As an example, he mentions the algorithm AlphaFold, developed with Google DeepMind. It can predict how proteins fold. This may be incomprehensible to most, but it has enormous significance for both life sciences and other key sectors.

“It has meant a quantum leap for biotech research. It can also be used for new green technologies, for example, better utilization of energy in waste. In the end, researchers will review the results and test them in the lab. But with the algorithm, you can much more quickly identify which solutions have potential and are worth testing in the lab. In this way, AI can accelerate the development of other new technologies and thus continue to lift the economy in the long term,” says Jens Lundsgaard.

That it is precisely Google DeepMind is a cue to what, according to Jens Lundsgaard, is absolutely crucial: Investments.

DeepMind was originally developed in the UK. But when capital was needed for further development, it ended up becoming part of American Google.

Top Executives Lag Behind

DeepMind is far from unique. In many cases, companies with immature but promising technologies have moved across the Atlantic when capital was needed. Over the past 15 years, American investments in information and communication technology have grown twice as fast as European investments, which is a crucial explanation for why Europe has fallen behind.

“Overall, the European economy is large, but as also pointed out in the Draghi report, there is not a unified, well-functioning capital market in Europe. It is absolutely crucial that we change this to ensure financing for growth companies in Europe,” says Jens Lundsgaard.

Besides investments, leadership stands as the other crucial steering signal that, according to Jens Lundsgaard, is needed to get Europe into a different AI gear.

This should not be understood as a desire for more engineers and computer scientists in the executive suites of large European corporations.

Boards and executive management should not have their hands deep in AI. But they need to better understand the business opportunities in AI. It is the top management that can secure investments and decide that the absolutely necessary development projects are initiated.

Unfortunately, there are figures indicating that it is precisely at the top executive level that understanding of AI’s potential lags the most.

A Deloitte survey looked at how Nordic companies work with generative AI. Of course, many companies are not yet involved, so the survey also asked about interest in the area.

While there is relatively high interest in AI among employees and line managers, only just under one in three top executives at so-called C-level—CEOs, CFOs, and others in actual corporate top management—express high interest in generative AI. Globally, far more top executives have understood that they need to be interested in AI. Here, six out of ten corporate directors have a high focus on the area.

It’s even worse at the board level in Nordic corporations. Here, only one in six of the surveyed board members has high interest in AI. Globally, it’s almost half.

“Considering that it is the executive management and boards that must make the absolutely crucial decisions, it is worrying that there is such low interest in the area. The decisions made in the coming years may be decisive for companies’ future competitiveness and market position,” says Jens Lundsgaard.

Slow Adaptation to New Technologies

There will be plenty of companies and entire sectors where the gains are not harvested as quickly as in the case of the AlphaFold algorithm. But typically, decisive new technologies accelerate labor productivity, and often this happens over a longer period.

In the note behind OECD’s estimate of AI’s impact on productivity, they looked at previous technological leaps such as the introduction of electricity, computers, and the internet. The crucial common denominator is that even though many may try out the new technologies, the penetration time can be quite long before the full productivity gains are achieved. This explains why OECD, compared to other researchers and think tanks, is cautious in its estimate of how quickly AI will affect productivity in the coming years.

Often, there is a period where the new technology is more or less used on the terms of the existing technology. Jens Lundsgaard mentions electrification as an example.

“At the start of industrialization, factories had several floors because there was a central steam engine, which via belt drive was the crucial energy source. For a period after electricity was introduced, factories continued to be built in the same design; only after some years did people realize that it was not necessary, as with several smaller electric motors, production could be organized differently. It was only then that factories began to be built on one level, where goods could be transported more easily, thus gaining the full benefit of the transition to electricity,” he says.

Jens Lundsgaard points to the story of electricity to emphasize that we still cannot foresee which changes AI will accelerate in the way companies organize themselves.

“There are many challenges, including some that concern transparency and ethics. We must solve these challenges while not hesitating to invest in AI. Because ultimately, it will be a decisive factor if we are to secure the prosperity we have in the Nordic and European countries,” he says.