Artificial Intelligence and Work

Events on the topic of Artificial Intelligence take place all over the country with titles such as "AI - supporting humans in solving global challenges" or "Automation and the 'Human Touch'." Such titles are characteristic and typical. The notion that AI will not replace but rather complement humans has become an omnipresent and dominant narrative today. Damn, I've explained this in my presentations before. Why? Because it's a reassuring thought. Unfortunately, it's also deeply misleading. If we want to effectively prepare for the impacts that AI will have on society in the coming years, it's important that we become clearer on this issue. It's understandable that people are receptive to a future vision in which AI primarily complements human activity. This idea allows us humans to maintain control and stay at the top of the cognitive food chain. It doesn't require us to make profound, uncomfortable changes regarding our place in the world. According to this line of thinking, AI is just another tool that we've cleverly created to make our lives easier, like the wheel or the internal combustion engine. But AI is not just another tool, and uncomfortable reorientations lie ahead of us. Chess provides a vivid example to start with. Once upon a time, playing chess was considered the epitome of human intelligence. In 1997, when the IBM computer program Deep Blue defeated chess world champion Garry Kasparov in a well-known duel on the second attempt, the defeated Kasparov famously said just before, "This match is a defense of the entire human race." What a humiliation! In response, the concept of Freestyle Chess or Centaur Chess emerged in the following years, where teams of humans and chess software competed against each other. The idea behind Centaur Chess was simple: While the best AI could now beat the best human in chess, an AI and a human working together (like a "centaur") would be the most powerful player of all, as human and machine would bring complementary skills to the table. It was an early version of the myth of augmentation. And indeed, mixed AI/human teams were able to outperform computer programs in chess for a while. "Centaur Chess" was praised as proof of the irreplaceability of human creativity. However, over the years, machine intelligence has exponentially advanced and left human chess players far behind. Today, no one talks about Centaur Chess anymore. AI has surpassed humanity in this area to such an extent that a human player would simply have nothing more to contribute. Chess is just a board game. What about in the real world? The myth of AI as augmentation has become widespread. An important reason for this: The prospect of job loss due to automation is a frightening one and let's say a very touchy subject in politics. Entrepreneurs, scientists, politicians, and others have a lot to gain by believing - and convincing others - that AI will not replace but complement humans in the workforce. Employment is one of the most basic social and political necessities in any society in today's world. To be openly job-destructive is therefore a losing proposition for any technology, company, and of course, any politician. "It's not about machines replacing humans, but about machines complementing humans. Humans and machines have different relative strengths and weaknesses, and it's about combining these two that will allow us to scale our goals and processes in the coming years 10-, 100-fold and beyond." Recently said a CEO, adding, "AI will bring humans and machines closer together." Three years ago, I was at a congress where the then IBM CEO Gina Rometti explained, "AI - could also be an acronym for 'augmented Intelligence,' meaning it would complement human cognition rather than replace it. Upon honest reflection, it becomes clear that many AI systems being built today will replace rather than complement large portions of the human workforce in the economy. The central promise of AI - the reason why we are researching it in the first place - is that it will be able to do things more accurately, more cheaply, and faster than humans can today. Once AI can deliver on this promise, there will be no practical or economic justification for the use of humans in many areas. For example, if an AI system can demonstrably drive a truck better and safer than a human under all conditions, there will simply be no sense in having humans continue to drive trucks. In fact, it would be harmful and wasteful to have a human at the wheel: Apart from the saved labor costs, AI systems never drive too fast, never get distracted, never drive drunk, and can stay on the road 24 hours a day without getting sleepy. The startups and truck manufacturers developing self-driving trucks today may not admit it publicly, but the goal of their research and development efforts is not to complement human labor. It's about replacing it. That's where the true value lies. The same applies in radiology. The primary task of radiologists is to examine medical images for the presence or absence of certain features, such as tumors. Pattern recognition and object recognition in images are exactly what Artificial Intelligence excels at. AI systems will not overnight replace humans, neither in radiology nor in any other field. Workflows, organizational structures, and people need time to change. The technology will not be perfect at the beginning. Initially, AI will indeed be used to support human radiologists: to provide a second opinion, or to review a variety of images and prioritize those deserving human review. And indeed, this is already happening. We are essentially in the "Centaur Chess" phase of radiology. But fast forward five years. Once it is undisputed that neural networks are superior to human radiologists in classifying medical images - across all patient groups, care situations, and disease states - will it really make sense to continue employing human radiologists? Particularly considering that AI systems will be able to instantly and at no additional cost review images of patients from anywhere in the world, and that these systems will constantly improve. Over time, the saying of AI legend and Turing Award winner Geoffrey Hinton will prove true: "We should stop training radiologists now. If you're working as a radiologist, you're like Wile E. Coyote in the cartoon; you're already over the edge of the cliff, but you haven't looked down yet." Meek meek What does all this mean for us, for humanity? A future where Artificial Intelligence replaces human activity rather than complements it will have a number of profound implications. I'll briefly touch on some of them, knowing full well that entire books could and have been written on these topics. First and foremost, job loss will be very painful and demeaning for people. This will happen across all social strata, geographical regions, and industries. From security guards to accountants, from taxi drivers to lawyers, from cashiers to stockbrokers, from court interpreters to pathologists - human labor will no longer be in demand in the economy, as increasingly sophisticated AI systems can perform these tasks better, cheaper, and faster than humans. Acknowledging this inevitability is not easy. Society must react quickly and creatively to cushion the impacts of this displacement of jobs. Sensible investments in retraining and requalification by the state and the private sector will be crucial. Even more fundamentally, a paradigm shift in how society conceives resource distribution will be needed in a world where material goods and services will become increasingly cheaper due to automation, while the demand for paid human labor will dwindle. The idea of a Universal Basic Income - incidentally developed in the mid-90s as the Ulm Model at the University of Ulm and until recently more of an academic thought experiment - is now being taken seriously by established political decision-makers. Last year, the Spanish government launched the largest basic income program in history. I believe that Universal Basic Income will become a normal and increasingly important political tool in the era of AI. An important realization about job loss due to AI is that some tasks will withstand automation much longer than others. Jobs where humans will be superior to machines for the foreseeable future are not necessarily the most cognitively complex. Rather, they will be those where our humanity itself plays a crucial role. These tasks primarily include those that require empathy, camaraderie, social interaction, activities that involve the "human touch." Babysitters, nurses, therapists, teachers, bartenders, and social workers will find work for many years to come. Likewise, people in roles that require creativity and unconventional thinking will not be replaced anytime soon. A clichéd but revealing saying about the relationship between humans and AI goes: While AI gets better at knowing the right answers, the most important task for humans will be knowing which questions to ask. Roles that demand this kind of resourcefulness include academic researchers, entrepreneurs, technicians, artists, and writers. In the jobs that remain over the years, people will spend less energy on boring, repetitive, soulless tasks and more on developing human relationships, managing interpersonal dynamics, and creative thinking. But make no mistake: Humanity is on the brink of a larger, more profound change as AI takes over more and more tasks currently carried out by humans. Put simply: We will eventually enter a world post-work. There will not be nearly enough meaningful jobs to employ every person of working age. More radically: We will no longer need people to work to generate the material prosperity necessary for the healthy livelihood of all. AI will usher in an era of prosperity. It will automate (and dramatically improve) the value-added activities that humans perform today; for example, it will enable us to produce food, shelter, and medicine synthetically on a large scale and at low costs. This is a stunning, almost incomprehensible vision of the future. It will require us to redefine our values and the meaning of our lives. Today, an adult's