Shortage of Data to Complement Shortage of AI Computing Power and Electricity in Europe

The European legislators want tech companies to seek prior consent before they can use data to train their artificial intelligence models. This will stifle startups, create a shortage of training model inputs and undermine European AI sovereignty.

Brian Williamson, a partner at the London-based Communications Chambers consultancy, predicts a ‘dark age’ for European artificial intelligence in his article posted on the website of the Center for European Policy Analysis (CEPA).

This is due to the fact that the European Parliament and several national governments consider rolling back text and data mining protections. Media and copyright holders argue for a shift toward a prior-authorization (‘opt-in’) default or mandatory statutory levies on AI training data. In April 2026, the French Senate even adopted a bill introducing a restrictive legal presumption that AI models will exploit cultural content.

To please media lobbyists and other copyright holders, European legislators overlook the fact that models do not copy or store data like a digital library, nor resell it. They learn from it, analyzing massive datasets of text, images, and code.

If an AI model generates an output that duplicates a copyrighted work, or arguably unduly exploits a creator’s likeness, copyright law should punish that output. But preventing a model from analyzing public text to understand grammatical structure or scientific concepts is the equivalent of banning a human art student from visiting a public museum.

If computers are restricted in what they can read and examine, this will stifle development and create high entry barriers. This threatens Europe’s competitiveness in three profound ways.

Firstly, it will preclude the development of startups and open source projects. If these have to obtain consent to use data, they will incur massive compliance and licensing costs. Large tech giants can afford armies of lawyers to negotiate these deals. European startups and open-source projects cannot.

Secondly, in addition to its energy and compute constraints compared to the US and China, Europe will face a shortage of data for training its AI models. Europe is already behind in AI, and local AI development will stall entirely if such rules are adopted.

Thirdly, it will undermine Europe’s sovereignty, as European models will be barred from learning from the open web, and enterprise, SMEs, and individuals will be unable to freely fine-tune European open models. Choice will be restricted, and Europe will become dependent on offshore models trained under more permissive legal frameworks.

Europe’s economic lag is natural now, as its output per hour grew by just 29 percent over the past three decades – versus 85 percent in the USA. AI has opened a window of opportunity to reverse this decline. Yet European policymakers risk slamming this window shut by tinkering with the legal architecture that permits quick and sound training of European-developed AI models.

Original publication: https://cepa.org/article/avoiding-a-european-ai-dark-age/