Germany struggles to keep pace in the AI race

German scientists are making significant contributions to research on artificial intelligence. But when it comes to translating that research into practical applications, Germany lags behind AI giants like the US and China — and that gap is growing. That's according to a new analysis by KfW, Germany's state-owned development bank. 

"With AI, as with many other areas of technology, we struggle to translate research into products that can be used by businesses," Fritzi Köhler-Geib, the KfW's chief economist and head of the bank's economics department, told DW.  

According to KfW, Germany imports a significantly larger number of AI-related products and services than it exports. This has made the country increasingly reliant on AI services from a handful of powerful companies abroad, raising concerns about its competitiveness in an increasingly AI-driven world. 

"When it comes to the implementation of AI, other countries are simply faster," said Köhler-Geib. "We have to make sure that Germany doesn't fall further behind." 

From labs to reality 

Work on artificial intelligence dates back to at least the 1950s. For a long time, it was seen primarily as "blue skies research" — scientific endeavors in domains where "real-world" applications are not immediately apparent.

But over the past two decades, advances in computing power and the development of novel techniques have brought AI technology out of the research lab and into society. In recent years, AI-powered programs such as ChatGPT have become everyday tools that people use to create text, images or even computer code.

 

However, despite significant contributions to the field by German scientists, bureaucratic hurdles and relatively low investment have hampered the progress of the German AI industry in developing real-world applications.

Germany 'lagging far behind' China, US in AI patents 

Today, the lag is evident in the number of new patents filed by German organizations, which is seen as a key measure of innovation. "When it comes to patent applications, Germany is lagging far behind China and the United States," said Köhler-Geib.

While China has seen a 100-fold increase in patent applications for AI technology since the early 2000s, Germany's growth has only tripled over the same period. Currently, Germany holds a meager 6% share of global AI patent registrations, far behind China's 29% and the US' 27%. 

"We import far more goods in this area than we export, while China, for example, has a significant export surplus in the field of AI," said Köhler-Geib, describing this trade deficit as "a significant weakness" in the race for AI. 

Germany unable to keep talent 

Her warnings were echoed by Alexander Löser, a professor at the Berlin University of Applied Sciences and Technology. When it comes to the machine-learning-based AI solutions dominating the market today, he said "Germany is increasingly becoming a customer for AI services, most of which are offered from outside of Europe — primarily the US, but also increasingly Saudi Arabia, Dubai and China." 

This trend is exacerbated by Germany's inability to hold on to some of its best AI talent, said Löser. "Many universities here do excellent research and train highly qualified people, but a significant number of them then choose to work abroad." 

At the same time, the strict regulatory environment in Germany and the European Union has put local companies at a disadvantage, said Löser, as they struggle to get access to what is often referred to as the "oil" that fuels most of today's AI systems: data. 

"Regulations are driving up the cost for our local AI ecosystem to acquire training data," he said. To reverse this trend, Löser said Germany should create open-source datasets for commercial use — "high-quality data that reflects our cultural values." 

In its report, Germany's KfW Bank also emphasizes the critical need for "adequate access to training data."

To catch up in the global race for AI, Germany also needs to boost investment in AI research and development, especially in areas where it already has a strong industry, and increase training opportunities for students and workers. 

Edited by: Rina Goldenberg