Toward an AI age "Wirtschaftswunder"

Post-war Europe paired automation with institution-building – and saw three decades of growth. With AI, we need to try once again

The prediction by Dario Amodei, CEO of Anthropic, that AI would soon create a “white-collar bloodbath” was one of the reasons why Markus and I started the Work/Code newsletter. We wanted to understand the impact of AI on the workforce better – and also how lawmakers, business leaders and professionals could shape it.

In a recent blogpost, Anthropic presented nine policy responses to mitigate the impact of AI on the economy, proposed by members of the company’s Economic Advisory Council and participants of the Economic Futures Symposium, a conference co-organised with Georgetown University.

The proposals range from investing into upskilling to new taxes on AI tokens and are a good starting point for this important debate – and yet they fall short of addressing the challenges that AI poses. New policies are important but not sufficient. We need to re-imagine our institutions as well: AI will not just affect our economy – it will also challenge our political system and the societies we live in.

So let’s take this wider perspective.

Transformational change needs supporting institutions

“Artificial intelligence isn’t just changing the job market”, essayist John Mac Ghlionn recently wrote in The Hill. “It is shattering the social contract that keeps nations stable.” In his view, citizens’ support of the political system is tied to the promise of advancement.¹

So how can we ensure that AI does not demolish but maintains the social contract – or maybe even reimagines it? After all, the legitimacy of governments was already strained over the past three decades through globalisation and political polarisation.

To answer this question, we need to look beyond policies and at institutions. Here are three ideas:

Build educational institutions fit for the AI economy

The three decades after World War II were not only characterised by massive automation, especially in Europe, but also the expansion of the education system which led to the creation of new jobs in the emerging knowledge economy.

As Carl Benedikt Frey et al. write: “While automation can provide a short-term productivity boost, it’s a one-time gain. If all progress since 1800 had been limited to automation, we would have efficient agriculture and cheap textiles, but little more”.

That is why governments also need to invest more in basic research and development which is the foundation of innovation. Basic technologies are often not developed on the free market, but in university labs.

AI can also help updating our educational system as ”Khanmigo” shows, an AI-powered tutor that helps students in their own pace. Instead of aiming at “superintelligence”, why not use AI to achieve ”mass intelligence” (Ethan Mollick)?

Learning does not end after graduation, and thus, skills development programmes as proposed in the Anthropic blogpost are of course essential. It is likely that this will primarily be a responsibility for governments (and there are some good examples), but governments can also enlist companies, i.e. by extending the amount of paid leave that companies have to grant to employees who do training.

Support the creation of complementary tasks, not automation

As Daniel Susskind writes, there are no “natural boundaries” for what AI can and cannot do, especially when we believe that we will at one point in time enter an age of ”superintelligence”.

But AI has limits, either in terms of efficiency (it might still be more efficient to use human labor for some tasks even if AI could do it), preference (such as art, sport, or teaching), and morals (e.g. judges, supervision). Similarly, it is plausible that AI may extend the market e.g. for software engineers or marketing professionals.

But using technology so that is does not lead to automation but is used to increase the productivity of workers and thus may even lead to wage increases depends on humans deliberately deciding on how to use AI, as Nobel prize winners Daron Acemoğlu and Simon Johnson point out:²

Most people around the globe today are better off than our ancestors because citizens and workers in early industrial societies organized, challenged elite-dominated choices about technology and work conditions, and forced ways of sharing the gains from technical improvements more equitably.

So institutions like collective bargaining that give workers a voice in how AI is deployed and how its benefits are distributed are essential.

The recent strike of Hollywood writers is an example of how this might look like: The agreement that ended the strike did not prohibit the use of AI, but prevents production companies from mandating the use of AI to the detriment of writers.

A good way to measure how evenly the benefits of AI are distributed is to look at impact of AI on the labor share.

Innovation and competition

Governments play a huge role in maintaining an innovative economy by protecting property rights and enforcing contracts.³ If we believe that AI will lead to a new age of discoveries (even if that may not happen quickly), a system that attributes the intellectual property rights of AI innovation in a way that sets the right incentives and does not create new monopolies will be a crucial element of the AI economy. We are only at the very beginning of this debate.

Likewise, the past wave of technological progress has led to numerous concerns about competition, the lock-in effects of major platforms and the emergence of ”superstar firms”. We know from past experience that monopolies have an adverse effect on innovation and welfare, so ensuring that AI does not create new monopolies or deepens existing barriers for innovation will be another central theme when building institutions for the age of AI. As AI reduces the barriers to entry in many industries, we should see a huge number of start-ups challenging incumbents, and not a further entrenchment of “superstar firms”.

Towards a new “Wirtschaftswunder”

Widespread automation is not unforeseen in history: After World War II, Europe automated large parts of its economy, starting with agriculture, and yet the post war years are known as the “Wirtschaftswunder” in Germany, “Les Trente Glorieuses” in France and “miracolo economico” in Italy. During these three decades of “economic miracle”, automation was accompanied by the creation of new jobs at massive scale which lifted the tide for all and increased welfare for the nation as a whole. And institutions played an important role in this development.

To repeat this “economic mircale”, we should not just look at whether our current policies are fit for purpose. We also need to re-imagine our institutions.

1. In our recent interview on the subject, Carl Benedikt Frey equally voiced concerns that the labor market impact of AI might lead to the creation of new coalitions and potentially a “white-collar luddite movement”.

2. Acemoğlu, Daron; Johnson, Simon (2023): Power and Progress: Our Thousand-Year Struggle Over Technology and Prosperity. New York: PublicAffairs.

3. Aghion, Philippe; Antonin, Céline; Bunel, Simon Bunel (2021): The Power of Creative Destruction: Economic Upheaval and the Wealth of Nations. Cambridge: Belknap Press.