In “The Scarcity of Labor”, Frankfurt sociologist Florian Butollo challenges the familiar notion that automation will leave us with less work to do. Instead, he argues, labor is becoming a scarce resource, a shift that is already fueling social tensions. He calls for rethinking how work is organized and taking a more forward-looking approach to AI, so it can help address the defining challenges of our time.
UniReport: Professor Butollo, your book presents a surprising thesis about work under capitalism, one that differs significantly from the conventional idea of ever-increasing rationalization. You describe a “dialectic of rationalization and expansion.”
Florian Butollo: Capitalism is synonymous with ongoing automation. Rationalization cuts costs and boosts efficiency, but it goes hand in hand with expansion – with rising demand and new needs. This translates into both more and increasingly complex products, new technologies and more complex value chains. It is not just fewer workers; instead, we are observing a shift in work: While working hours may be reduced in some areas, the workload in others will increase because of rising demands, the need for new skills and the greater interconnectedness of work. The trend points towards more work, although it is important to take a differentiated approach: A U.S. study, for instance, found that 60% of the occupations people work in today did not exist in 1940.
So, automation and AI are also creating new jobs?
Exactly. Software development is one example. While new tools may reduce the amount of work involved, the demands placed on them are exploding. We are faced with more applications, larger volumes of data and more demanding infrastructures. The result is that there are far more software developers today, roughly half a million more than just 15 years ago, despite ongoing automation.
What role does AI play?
AI acts as a catalyst in a system that is already driving rationalization. It shifts workloads and changes what is expected of workers, in software development and customer service, for instance. However, AI does not simply replace entire professions. It changes tasks, increases pressure to keep developing new skills and creates new areas of work where expertise, sound judgment and the responsible use of AI are essential.
One widely held view is that automation will mean we work less in the future and might use the time gained in other ways. What is your view?
This is a debate that goes back to John Maynard Keynes and continues today. Reality is more nuanced: In many fields, the volume of work is increasing despite automation. Demographic change adds to the pressure: There are fewer workers available, while demand is holding steady or rising. External pressures, such as competition, speed and market share, leave less room for learning. Without time for experimentation and adequate opportunities to build skills, workers risk becoming frustrated and overwhelmed.
Anyone who has experienced the introduction of digital communication tools at work is familiar with the pattern: At some point, we no longer had to handwrite letters, type them on a typewriter and send them. Phone calls were partly replaced by email and messaging services. And while communication has become both easier and faster, we are now faced with a constantly growing flood of digital data and messages. Is this another example supporting your central thesis?
Absolutely. The free time that simpler, more efficient communication was supposed to give us is nowhere to be seen. Few people would say that communication has become less of a burden in their working lives. Chat formats have amplified email communication, but they also demand even more rapid-fire attention. If anything, communication has become more intense.
This is surely especially true in the service sector. But what about manufacturing, for example, in the automotive industry that is so important to Germany?
The automotive industry has become extraordinarily automated over the past decade. It has high-tech production lines, lights-out factories in its paint shops and much more complex electronics. Yet employment has not declined globally. Why? Because there are more models, more production changes, more variations in product details and greater engineering demands. Automation reduces routine work but increases the need for staff in development, software, maintenance and logistics.
The German automotive industry is currently losing a lot of jobs.
The crisis in Germany’s automotive industry is both real and severe and serves as a stark counterexample to Germany’s labor shortage, because tens of thousands of jobs are at risk. At the global level, however, the trend is toward more work: OECD countries have added millions of jobs since 2010, and employment is rising worldwide despite automation. The underlying finding remains unchanged: Work is expanding in many fields, even as some are shrinking. Incidentally, Germany’s automotive industry today employs more people than it did in 2010.
How would you characterize the German labor market?
Employment growth in Germany’s health and social services sectors is more than offsetting losses in industry. Demographic change is increasing demand for care and health services, which is creating more jobs in those areas. At the same time, there is a risk of deindustrialization. My central argument is not that capitalism inevitably leads to a growing volume of paid work. It is that there is a structural labor shortage because fewer people are available while demand in the labor market remains high.
Your book emphasizes the political dimension of technological development. What can be done to ensure AI has positive effects on society?
When labor becomes a scarce resource, we have to ask which kinds of work should be prioritized and which we might do without. We need to consider how incentives can encourage work in fields that matter to society. That calls for a different approach to innovation and investment policy. In her book “The Entrepreneurial State”, economist Mariana Mazzucato shows that all key technologies used in the iPhone, including the touchscreen and voice recognition, were byproducts of government investment in high-risk areas connected to military purposes. Public innovation policy could similarly support work in important future fields, such as climate, health and infrastructure, if it puts the public interest first.
Many people fear that technology is driving environmental destruction. Is AI adding fuel to the fire or does it present a solution?
In its current form, AI is more likely to add fuel to the ecological crisis, because it consumes immense resources. Sensible alternatives require new political priorities, including a fair allocation of resources, regulation and public investment in projects that serve society. Right now, an incredible amount is being invested in getting products to market faster, as in ultra-fast fashion. That ties up huge amounts of labor in an industry that fuels a harmful cycle of growth and acceleration, both economically and ecologically. The way AI is being used fits seamlessly into this pattern.
You also turn your attention to academia itself. Do researchers still have room to pursue bold ideas at a time of great complexity and intense pressure to publish?
Yes, but the pressure is immense. As the saying goes, “publish or perish.” We see constant measurement of research output. Still, there is room for big, ambitious research questions, even if they are harder to pursue. Whether there will be enough room in the long term depends on institutional structures, education policy and research funding. The scope for such work is shrinking as funding faces dramatic cuts.
What kinds of responses to your book have you heard personally, from fellow academics as well as practitioners?
They have been very varied. Many people recognize the dynamics I describe: more work, a faster pace and higher demands. Others warn against pessimism. Overall, the discussion seems to acknowledge that work is changing rather than disappearing, and that learning and transformation are necessary. One striking response has come from practitioners, including software developers, who confirm that despite the scale of cloud computing, workloads and customer demands are growing enormously.
What specific recommendations would you offer policymakers, businesses and academia?
In general, we should take a broader view: In addition to efficiency, we also need to consider how work and resources are allocated. As I said, we must create room for learning and experimentation so that innovation does more than cut costs and can also address social problems. I want the debate to avoid both apocalyptic thinking and uncritical optimism. Instead, it should lead to a realistic, historically informed discussion about what we want to work for and how work can be allocated sensibly in an increasingly complex society.
Interview: Dirk Frank

Florian Butollo

Florian Butollo