Artificial Intelligence occupies a curious space in European policy discussions. Its position as a laggard in the AI “race” has become a symbol for the continent’s real and perceived lack of competitiveness and digital sovereignty, especially compared to the US and China. At the same time, AI is also hailed as a solution for a whole range of complex problems the continent is facing: the climate crisis, slowing economic growth, and deteriorating public services.

Enter, Europe’s nascent industrial policy on AI: strategies to guide government spending, investment, and regulatory strategies focused on the AI industry. After the DSA, the DMA and the AI Act have been adopted or entered into force, we expect that the next Commission’s main focus will be on digital industrial policy (and enforcement). This is part of a broader turn towards economic statecraft we have documented globally (see our previous work here)

Europe’s newfound interest in industrial policy comes with opportunities and unintentional dangers. Instead of creating a vibrant innovation ecosystem in Europe, industrial policy could end up further entrenching the power of Big Tech, who already exercise a perhaps irreversible grip over the global AI ecosystem through control of capital, research and infrastructure. What is often misunderstood in Europe, is that concentration in the tech industry is constitutive of the AI market as we know it today. As a result, policymakers that are invested in developing a more competitive homegrown AI market, place industrial policy measures in concert with anti-monopoly, privacy and other policy goals.

As public funds begin to pour in for AI industrial policy, European lawmakers must confront the elephant in the room: as it stands today, there is no AI without Big Tech. While European AI startups espouse aspirations to compete, they must contend with the reality that large US and Chinese-based tech companies control core AI infrastructure for both development and deployment: compute, data, talent, and most importantly – access to the consumer.

In order to confront these infrastructural and ecosystem dependencies, we must start the crucial work of imagining an alternative digital future for Europe that includes an innovation paradigm outside of the incentive structures set by these companies. This must also stretch beyond the narrow imaginary of what counts as “public good” and “public interest” in current EU debates and include a more sophisticated understanding of the role that digital technology can play in innovation and economic growth more broadly. This will involve distancing from the current hype-cycle of chatGPT-inspired AI use cases, the continual push to build AI at ever larger scale (necessitating access to resources these firms control) and the geopolitical impulse towards the so-called “AI arms race”. 

If we don’t, efforts towards industrial policy on AI may eventually end up strengthening, rather than contesting, AI monopolies. And instead of empowering citizens, workers, and the environment, blanket adoption of AI could lead to waste, rights abuses, and single points of failure. We need only look to the Dutch child benefits scandal, or more recent evidence about a dysfunctional automation system used by Danish child protective services, for examples of how AI adoption can go horribly wrong, automating austerity measures, disenfranchising communities of color and the poor, and leading to significant societal harm. 

Building on our previous work on AI Nationalism, Computational Power, and Antitrust, and in collaboration with a network of researchers and advocates, AI Now will be closely monitoring Europe’s nascent industrial policy on AI. Over the coming months/year we will interrogate the assumptions, central tensions, and trade-offs at the heart of Europe’s desire to become a major player in AI.

Among other things, we will start by provoking the following questions: 

What would it take to realise a European vision for public interest AI? Given that current market structures heavily favor today’s tech giants, is a European vision for AI even feasible? 

Can Europe ever be competitive, or reach its goals on sustainability and climate change within the current paradigm of building ever larger models? 

What does public interest mean, when the AI Act failed to protect the rights of migrants and others who are most vulnerable? 

And given that so many benefits of AI remain speculative – should Europe even invest in large scale AI at all? 

With a network of collaborators and experts from across the region and beyond, we hope to start moving from diagnosis to an actionable agenda to harness the opportunity before us.  

How Europe wrestles with these questions is vitally important – not just for the continent. If a continent as wealthy and prosperous can’t truly compete or build public digital infrastructure and goods – this doesn’t bode well for regions that already find themselves at the global economic margins. 

We look forward to sharing more from this work as it moves forward, and welcome insights along the way.