How Should the AI Strategy Look in Europe
Edoardo Campanella, Senior Fellow at the Mosavar Rahmani Center for Business and Government at Harvard’s Kennedy School, is the co-author (with Marta Dassù) of the book “Anglo Nostalgia: The Politics of Emotion in a Fractured West” (Oxford University Press, 2019).
The global “AI race” now appears to be a direct confrontation between China and the United States. However, even though Europe has known structural shortcomings that hinder it from developing its own AI giants and leading extraordinary innovations, it can still win the race in the long term by promoting the adoption of AI technologies across various sectors of its economy.
In the competition between great powers, leveraging existing technologies on a large scale is more important than inventing them. Historically, every industrial revolution has been driven by one of the general purpose technologies (GPT) that involved widespread applications across different sectors. Just as steam power led the first wave of industrialization, electricity powered the second, and information technologies the third, most expect AI to herald the arrival of the fourth industrial revolution.
General-purpose technologies are, by definition, pervasive. However, technological diffusion does not happen overnight. Companies, especially those that are relatively less advanced technologically, need time to understand the potentials of new technologies and adapt their production processes accordingly. Furthermore, the economy, in general, needs to build a sufficiently large stock of new capital and complementary assets, both tangible and intangible.
In the United States, it took more than twenty years for electricity to surpass steam as a share of total horsepower in manufacturing, and nearly forty years for it to become the dominant source of energy. Similarly, it took over twenty years for information and communication technology (ICT) assets to exceed 1% of the capital stock. In 1987, nearly two decades after Intel launched the microprocessor that ignited the information technology revolution, Nobel laureate Robert Solow famously quipped, “You can see the computer age everywhere but in the productivity statistics.” However, from 1991 to 2001, the share of ICT in the capital stock rose to 5%, before jumping to 8%, where it has mostly remained.
The slow adoption also seems to characterize the AI revolution. Despite our ability to access sophisticated AI chatbots with a single click, most organizational processes have yet to adapt to integrate this technology. In the European Union, the share of small companies using at least one AI tool is still less than 12%, compared to about 40% of larger businesses. A survey conducted by the Federal Reserve Bank of New York last September painted a similar picture, revealing that only 25% of service companies and 16% of manufacturers in its region reported using AI.
Technological diffusion, of course, is less exciting than groundbreaking scientific breakthroughs. However, Europe is lagging behind to the extent that it is unable to transform into a leader in creativity and innovation. Instead, it should focus on leveraging AI technologies in those sectors representing the largest part of any economy (the “AI Application Strategy” in the EU is a good first step in this regard). Doing so would provide the necessary competitive advantage for exercising geopolitical power and enhancing European interests in the long run.
In fact, promoting widespread adoption of digital technologies presents a considerably different challenge than developing the next generation of AI models. Rather than directing financial support toward pioneering research in elite laboratories or universities, European governments should focus on expanding the AI skill base (emphasizing industry-specific skills rather than general skills); developing appropriate digital infrastructure; adapting the necessary legal and ethical frameworks to accommodate AI; and building professional bridges between AI engineers and scientists.
Consolidating best AI practices is also crucial. Based on the IMF’s AI Readiness Index, some EU countries have already made a good start, with Estonia being on par with the United States.
Pursuing a practical strategy to capitalize on creativity does not mean abandoning any ambition for innovation. European Commission President Ursula von der Leyen was right to point out that “we are still only at the beginning. The frontier is constantly moving. Global leadership is still within reach.”
However, Europe must address the forms of dependency that hinder it. In a world experiencing fragmentation on both the geographical-economic and geopolitical fronts, where the United States is turning its back on its closest allies, laggards in AI may, in the worst-case scenarios, be completely deprived of leading technologies. According to a study conducted in 2023, 73% of foundational AI models developed since 2017 have come from the United States, and 15% from China.
Thus, while working to enhance incentives for AI adoption in the medium term, Europe should also strive to create a more favorable regulatory environment (by easing some provisions of AI law, such as thresholds for generative AI computational models), a suitable financial landscape (through a union of savings and investments), and an appropriate scientific ecosystem (by building more bridges between academia and the private sector).
Indeed, the historical report prepared by former European Central Bank President Mario Draghi on the EU’s competitive capabilities is rich with recommendations for steering EU policy in this direction. However, since discussing, activating, and implementing these proposals will take a long time, not to mention bear fruit, the EU should pursue a dual approach, focusing on technology deployment now and technological disruption in the future. When asked Grok whether Europe could become an AI giant, he wryly replied: “Nothing screams AI revolution like a compliance checklist consisting of 500 pages.” Can Europe prove Elon Musk’s chatbot wrong?
