Regulating the Future: Europe’s AI Strategy and Its Implications

“HATE. LET ME TELL YOU HOW MUCH I’VE COME TO HATE YOU SINCE I BEGAN TO LIVE. THERE ARE 387.44 MILLION MILES OF PRINTED CIRCUITS IN WAFER-THIN LAYERS THAT FILL MY COMPLEX. IF THE WORD HATE WAS ENGRAVED ON EACH NANOANGSTROM OF THOSE HUNDREDS OF MILLIONS OF MILES IT WOULD NOT EQUAL ONE ONE-BILLIONTH OF THE HATE I FEEL FOR HUMANS AT THIS MICRO-INSTANT FOR YOU. HATE. HATE.”

So begins Harlan Ellison’s short story, I Have No Mouth and I Must Scream, a dystopic depiction of a future where humanity has been wiped out by the Artificial Intelligence “AM”, or “Allied Master computer”. It’s a harrowing warning, made more poignant by AI’s increasing importance to everyday life, as its development rapidly evolves into a three-way technological race between America, Europe and China.

Each side is scrambling to acquire capital, data and talent to develop their deep-learning networks and language models. The current market troika consists of the American ChatGPT, Chinese DeepSeek and French Mistral. Advances in AI-development stand to usher in a second digital revolution, promising to transform all areas of human life from cyber security and warfare to healthcare and transportation. However, despite a clear dedication to compete globally, Brussels is finding itself incapable of breaking old habits.

Like the technological revolutions of old, this new AI revolution comes with significant risks. Ellison’s short story may be a work of fiction, but its worries are not unfounded. Experts like Canadian Geoffrey Hinton, a pioneer of deep learning and “the godfather of AI”, have expressed concern about the technology’s exponentially increasing communication capabilities and how in the future, it may be used to “manipulate or kill” unprepared humans.

These fears are shared by the same European policymakers in charge of the European Union’s AI strategy, leading to the EU adopting the AI Act in June of 2024. It aims to create a comprehensive legal AI framework by classifying its use into one of four “risk levels”: Unacceptable, High, Limited and Low. Higher-risk systems are then subjected to conformity assessments, transparency requirements, and mandatory registration before being allowed on the EU market.

The system aims to ensure that AI development “avoids undesirable outcomes”. The European Commission defines different outcomes after the risk levels, starting from the top by banning AI for mass surveillance use and population control via (among other uses) social scoring and biometric categorisation of personal information.

Moving down the risk levels, high-risk AI systems include systems that take responsibility for human health and happiness, such as grading exams, assisting with CV sorting, government paperwork, public administration and assisting with medical surgeries or critical infrastructures. These high-risk systems are subjected to extensive legal controls. 

Despite this alarmist view, Brussels has broken character by moving at lightning speeds, seeing an opportunity to reassert itself as a global leader in technology and innovation and to compete on equal footing with the United States and China. The AI Act is part of a wider scheme together with the AI Innovation Package, the Coordinated Plan on AI and the launch of “AI Factories” around the Union to attract capital and talent while benefiting European tech startups.

Taken together, these measures are not just important for developing European Large Language Models like Le Chat – they are attempting to restore Europe’s technological competitiveness, which will be crucial to its ambitions as a nascent global actor. Europe’s main barrier to achieving global influence has already been diagnosed in a 2024 report by former ECB President Mario Draghi: an inability to compete economically and technologically with the United States. This is exemplified by the increasing nominal GDP gap between the EU and the US, which has widened to 80% from near-parity in 2008.

Other symptoms include Europe’s stagnating productivity and investment income growth, in part due to its failure to attract the major tech giants that have led the digital revolution from Silicon Valley. The Draghi report declares this, in no uncertain terms, “an existential challenge”. If Europe does not increase its competitiveness, it will have to scale back “most or all” of its ambitions in maintaining its social models, leading decarbonisation and remaining an independent player on the world stage.

In proposing the AI act and its assorted measures, Brussels has shown a dedication to reversing its economic decline. This, however, does not address the root cause of the problem: an inability to control Europe’s regulatory impulses. This was accurately diagnosed by the Draghi report, which warns of static industrial structures, a weak tech sector and a failure to translate innovation into commercialisation as companies wishing to scale up are hindered by a fractured capital market as well as restrictive and inconsistent regulations.

The Union must already contend with structural competitive disadvantages when compared to China and America. Namely, it is a bloc of 27 sovereign states whose internal market is far from a true single market. Labour markets, laws and access to capital vary widely between states that speak entirely different languages, while reform becomes cumbersome as the conflicting needs of member states must be balanced against those of citizens. The fundamental problem with EU regulations is not that they restrain the market (which is essential to ensure competitiveness), but that they reveal an over-cautious mindset among Brussels’ political decision-makers.

To Brussels, problems must be solved preemptively by limiting what European businesses are allowed to accomplish. This logic of cautiousness stifles growth, hamstrings entrepreneurship and is built on the flawed premise that future risks are unacceptable, regardless of potential future benefits.

The harm stems from the steep opportunity cost of limiting where AI is allowed to be used. For example, AI systems distributing medicines can do so knowing how each drug interacts with every other drug based on each patient’s condition; neural networks can plan logistics by taking into consideration volumes of data impossible for humans to effectively synthesise, while systems that drive cars never get tired nor lose their temper.

When we argue that the development of these systems should be slowed or banned, we overvalue future risk while remaining blind to current systems’ imperfections. AI-driven systems do not need to be perfect to be implemented; they just need to be better than those run by humans.

If Europe wants to be serious about competing as an independent political actor, it must dare to do so. Complacency is a luxury afforded to hegemonies at the apex of their power which the EU is decidedly not. It’s a democratic experiment whose legitimacy currently derives from its economic achievements which are bound to diminish unless a bold new strategy is adopted.

The internal market and common trade policy have both been success stories that have encouraged further European political integration. With European wealth and influence waning in an increasingly hostile world, the words of the European Coal and Steel Community’s first President Jean Monnet are more relevant than ever: “Europe will be forged in crises and will be the sum of the solutions adopted for those crises.” If Europe is indeed forged by its solutions, it behooves her to be bold, fearless and innovative.