The global “AI race” now seems to have been reduced to a head-on clash between China and the United States. But even if well-known structural deficiencies are preventing Europe from developing its own artificial intelligence ( AI ) giants and pioneering seminal innovations, it could still win the race in the long term by promoting the diffusion of AI technologies across its economy.

In a great-power competition, leveraging technologies at scale is more important than inventing them. Historically, each industrial revolution was propelled by a general-purpose technology ( GPT ) that had wide-ranging applications across sectors. Since the steam engine drove the first wave of industrialization, electricity the second, and information technologies the third, most expect AI to usher in the fourth industrial revolution.

A GPT is pervasive by definition. But technological diffusion does not happen overnight. It takes time for firms, especially those that are relatively less technologically advanced, to understand a new technology’s potential and adapt production processes accordingly. Moreover, the broader economy needs to build a sufficiently large stock of new capital and complementary assets, both tangible and intangible.

In the US, it took more than 20 years for electricity to surpass steam’s share of total horsepower in manufacturing, and almost 40 years to become the undisputed source of power generation. Similarly, it took more than 20 years for information and communications technology ( ICT ) equipment to exceed 1% of the capital stock. In 1987, nearly two decades after Intel launched the microprocessor that gave rise to the IT revolution, Nobel laureate Robert Solow famously quipped: “You can see the computer age everywhere but in the productivity statistics.” But then, between 1991 and 2001, the ITC share of the capital stock rose to 5%, before jumping to 8%, where it has roughly stabilized.

Slow adoption also seems to be a feature of the AI revolution. Even though one can access sophisticated AI chatbots with a click, most organizational processes have not been adapted to integrate the technology. In the European Union, the share of small enterprises that use at least one AI tool is still below 12%, compared to around 40% of larger companies. A survey by the US Federal Reserve Bank of New York painted a similar picture last September, finding that only 25% of service firms, and 16% of manufacturers, in its region reported using AI.

Of course, technological diffusion is less exciting than pioneering scientific breakthroughs. But Europe is too far behind to become a cutting-edge innovator. It should focus, instead, on leveraging AI technologies in those sectors that represent the largest chunk of any economy ( the EU’s Apply AI Strategy is a good first step ). Doing so would confer the competitive advantage needed to exercise geopolitical power and advance European interests over the long term.

Promoting the widespread adoption of digital technologies is a markedly different challenge than developing the next generation of AI models. Instead of channelling financial support toward frontier research in elite labs or universities, European governments should focus on widening the AI skills base ( emphasizing industry-specific skills, rather than general ones ), developing the appropriate digital infrastructure, adapting legal and ethical frameworks to accommodate AI, and building professional bridges between AI engineers and scientists.

Standardizing AI best practices is also crucial. Here, judging by the International Monetary Fund’s AI Preparedness Index, some EU countries are already off to a good start, with Estonia on par with the US.

Pursuing a pragmatic strategy to capitalize on innovation does not mean abandoning any ambition to innovate. European Commission President Ursula von der Leyen is right to point out that, “We are only at the beginning. The frontier is constantly moving. And global leadership is still up for grabs.”

But Europe will need to address its dependencies. In a world undergoing geoeconomic and geopolitical fragmentation, with the US turning its back on close allies, AI laggards could – in the worst-case scenario – be deprived of frontier technologies altogether. According to a 2023 study, 73% of AI foundation models developed since 2017 came from the US, and 15% from China.

So, while strengthening incentives for AI diffusion in the medium term, Europe should also create a more accommodating regulatory environment ( such as by relaxing some provisions of the AI Act, like computational power thresholds for generative AI models ), financial landscape ( such as through a savings and investments union ) and scientific ecosystem ( with more bridges between academia and the private sector ).

Former European Central Bank President Mario Draghi’s landmark report on EU competitiveness is full of recommendations to move EU policy in this direction. But since these proposals will take time to debate, enact, and implement, let alone to bear fruit, the EU should pursue a two-pronged approach, focusing on technological diffusion now and on technological disruption in the future.

When I asked Grok whether Europe can become an AI giant, it sarcastically responded: “Nothing screams ‘AI revolution’ like a 500-page compliance checklist.” Can Europe prove Elon Musk’s chatbot wrong?

Edoardo Campanella is a senior fellow at the Mossavar-Rahmani Center for Business and Government at the Harvard Kennedy School.

Copyright: Project Syndicate