#42 State of AI Education
How does Europe compare to the US when it comes to the size of the AI workforce?
When looking at the landscape of AI, it's common to assume the United States holds a significant lead over Europe across all AI domains. However, Europe's AI workforce is substantial and comparable to the US. For example, according to Understanding Recruitment, the number of machine learning engineers, data scientists, and data engineers in Europe matches that of the United States in 2023. Additional data from the Center for Security and Emerging Technology (CSET) reveals that around 6 million individuals are actively involved in AI system development in the U.S., translating to approximately one in 26 jobs. Sequoia's "Atlas of AI" provides further perspective, estimating approximately three million technical talents in Europe. With a workforce of 199 million, this suggests roughly one in 33 jobs in Europe can be linked to AI. These figures closely align with CSET's findings, indicating a minimal difference in the skilled AI workforce between Europe and the U.S. The differences between the USA and Europe can be found elsewhere. For example, the US has significantly more money for the development of AI systems and is home to more top researchers than Europe.
Innovation happens when AI talent comes together in large companies
AI skills are not evenly distributed across an ecosystem. Rather, they are concentrated in a few companies and tech regions that are able to reap the benefits of the technology. The touted benefits are manifold: AI promises to speed up the production of goods, improve the quality of communication through Large Language Models (LLMs) or optimize routes for trucks. However, one of the most important capabilities of AI is to drive innovation. This is the result of a study by Tania Babina and her team, which was published this year in the Journal of Financial Economics. To come to this conclusion, they analyzed more than 535 million employee CVs and 180 million job advertisements, which gave them an indication of the level of AI skills in companies. Larger companies in particular can use AI to improve their brand, patent new products and update the company's portfolio. One explanation for this effect is that AI makes experimenting with products easier and faster. The faster learning process encourages companies to dare to update their product portfolio. In this sense, AI in large companies does something very similar to GenAI tools for knowledge workers: it encourages idea generation.
To what extent have US companies adopted AI?
AI adoption remains concentrated across the corporate landscape. A study conducted in 2022 by Kristina McElheran and her team, analyzing over 850,000 firms, revealed that only 6% have deployed AI systems into their operations by that time. Larger enterprises with over 5,000 employees are leading in AI implementation, while smaller companies show slower uptake rates. The study also emphasizes the reliance of AI on enabling technologies like cloud computing. It found that the vast majority of AI adopters had previously undergone digitization and integrated cloud solutions, indicating the need for robust data infrastructure for AI deployment. A key finding is the potential for an AI divide, where larger corporations gain disproportionate access to advanced AI capabilities. This highlights the importance of addressing disparities in AI adoption across companies.
New report by the Stifterverband on AI innovation in Germany
A recent publication by the Stifterverband attempts to find an answer to the question of how the innovative potential of AI can be promoted in Germany. I took a close look at what the report says about what needs to be done in terms of AI education. From page 27 to 32, each page deals with one aspect of AI education. One thing that struck me was that Germany has the highest proportion of STEM students in universities across Europe, 38% in 2021. The proportion has remained constant since 2001. Looking at schools, the number of hours pupils spend on IT varies across Germany and is still at a fairly low level. In Bavaria and Baden-Württemberg, for example, pupils receive 1 to 2 hours of compulsory IT lessons. Mecklenburg-Vorpommern is the frontrunner with 5 to 6 hours per week. Those already working in AI-related jobs are very well educated, with 84% of them holding a Master's degree or higher. The final topic they address in their analysis is a potential "brain drain" of AI talent in Germany. Their data comes from a study conducted by Pegah Maham in 2023, and while this study suggests an outflow of professionals with PhDs, other studies looking at the overall AI workforce report a positive influx of talent into Europe and Germany. Overall, the data presented in the report suggests that Germany has a solid foundation for building a strong AI workforce. The biggest challenge is not the size of the workforce, but attracting top talent to Germany.
Duisburg receives 18 million for the establishment of an AI center
Since Germany decided to phase out coal, some regions in Germany, particularly in North Rhine-Westphalia, might face economic problems. In order to help these regions remain competitive, the German Bundestag launched the "5-Standorte Programm" in 2020, which includes subsidies of 1.09 billion euros until 2038. The city of Duisburg has now given the go-ahead for the “Zentrum für angewandte Künstliche Intelligenz in Duisburg” (ZaKI.D). The 17 million euros invested in the program will be used to create educational formats on the topic of AI and a new incubator program for start-ups. The center will focus in particular on embedded systems and the use of AI with sensor data.