LBBW Research: How AI will change work
Artificial intelligence is changing the world of work. New occupations are emerging, old ones are dying out. But how will AI affect employment?
Science-fiction movies have been playing with the idea for decades: A human hero who takes on an intelligent machine. Hollywood star Harrison Ford did this in 1982, when he hunted down human replicants in “Blade Runner,” just like the man-machine hybrid “The Terminator” in the guise of Arnold Schwarzenegger a decade later. At the turn of the millennium, Keanu Reeves also had to battle artificial humans in a digital parallel world, the Matrix. These apocalyptic imaginings all remained cinema fiction. Nonetheless, they fanned fears that artificial intelligence (AI) could increasingly influence reality, especially employment, with robots taking on ever more sophisticated duties and thus supplanting humans and wiping out jobs.
Artificial intelligence (AI) is when human perceptions and human self-determined actions are successfully emulated by machines. Such systems are playing an ever greater role; they are even increasingly shaping our everyday lives. As long as it provides useful assistants like Apple’s Siri, Amazon’s Alexa and Google Now, new technology is welcome because it makes life more comfortable. The world’s major tech companies are betting their futures on it. In Seoul, South Korea, public life came to a standstill a while ago when the country’s best Go player lost the popular board game to a computer. Go hotshot Lee Sedol stood no chance against the learning neural networks. Optimists expect AI to surpass even the internet in significance and claim it compares only to the invention of electricity. Intelligent robots are on the march – bringing with them the question of how they will affect the world of work.
One in five companies in the digital industry sees artificial intelligence as a top issue, according to Germany’s digital association Bitkom. In 2016, this figure was just 9%. “When companies deploy AI, they hope for higher added value,” says Dr Mathias Weber, head of IT Services at Bitkom. At this year’s Hannover Messe, the consulting firm Sopra Steria Consulting asked corporate decision-makers for their thoughts. Across industries, nearly 80% expect the use of AI to have positive effects on staff costs and personnel expenses in production. However, this expectation is not so much associated with the prospect of replacing human work with robots. In automotive production, for example, it is rather about considerably reducing monotonous and dangerous work with intelligent assistant systems. According to the companies surveyed, therefore, AI does not deserve its reputation as a job killer. Raúl Rojas, Professor of Artificial Intelligence at the University of Berlin, has a more pessimistic view of the situation: “The effects on employment in the next few years will tend to be negative. I don’t see how so many people can be retrained in such a short time,” he says.
“The biggest problem will not be that we run out of work to do, but that the nature of the work will change and we will have to adapt quickly,” says Dr Guido Zimmermann, Senior Economist at Landesbank Baden-Württemberg (LBBW), who has examined the issue in depth recently. “Curiously, the technological breakthroughs that artificial intelligence can bring about are overestimated in the short term. However, its long-term impact is significantly underestimated,” says the LBBW economist.
The LBBW study shows that some industries are no longer conceivable without AI. The financial sector, for example: FinTech companies are springing up like mushrooms, developing digital consultants or apps for personal financial planning. AI is also the foundation for the much-discussed autonomous driving in the automotive industry. And in manufacturing, process automation is optimizing supply chains, and production on demand is increasingly common. Business models that don’t draw on AI methodology are hard to imagine in the medium to long term. “Corporate structures will become more virtual, the significance of production and sales platforms will increase,” says Dr Zimmermann.
Overall, Bank of America estimates the market volume for robots at around USD 67bn by 2025. The associated gains in efficiency could generate an annual effect of USD 14tn to 33tn up to 2025. These rough estimates show that, in the long term, new AI technologies are expected to result in higher labor productivity and thus higher real wages for workers. However, the conditions are not the same for everyone. Some sectors will win, some will lose, new occupations will be created, others will die out, and jobs will change. Professor Raúl Rojas forecasts very good prospects for highly educated workers like engineers. “But the middle is collapsing; in the USA, the proportion of industrial workers is now just 10%,” says the digital thought leader.
In the long term, a study by Landesbank Baden-Württemberg comes to a different conclusion: With robots becoming ever smarter and taking on non-routine jobs, the current polarization of the labor market (“middle-skilled jobs are disappearing, low- and high-skilled jobs are growing”) will increasingly give way again. The winners are likely to be employees who know how to use AI methods or who do jobs that cannot be done by a robot.
So far increasing automation has primarily disadvantaged workers who carry out routine tasks. This could be politically explosive, as the economic basis of the middle class of developed nations has previously been based chiefly on repetitive tasks. Economist Dr Zimmermann assesses the trend as follows: “We suspect that AI should tend to be positive for society, in the knowledge that the economic and social shocks caused by these new technologies will have to be cushioned by legislation.”
„In the long term, the impact of AI is significantly underestimated.“
Dr. Guido Zimmermann, Senior Economist, LBBW
Dossier: Potential and Risks of Digitalization
Dossier: Our analysis states the challenges of the digital transformation and sketches out approaches to a solution for businesses and the state.