How to regulate AI – broadly or narrowly?

The European stance

In recent years the European Union has taken a pioneering stance in shaping the regulatory landscape for AI. The cornerstone of this endeavour is the proposed Artificial Intelligence Act, a framework designed to establish harmonised rules for AI development and deployment. A defining feature of the EU's approach is the adoption of a risk-based paradigm, which encompasses regulating the various applications and functionalities of AI systems rather than the technology itself.

This strategy views technology as a versatile tool with the potential to be either beneficial, neutral, or detrimental to society depending on how it is applied. This approach – which might be referred to as narrow regulation – concentrates on pinpointing potential benefits and hazards.

From narrow to broad

The political landscape for AI regulation shifted in 2023 with the emergence of powerful generative AI models capable of producing intricate works of art and sophisticated linguistic content. This triggered a heated debate about the adequacy of narrow regulation to mitigate the associated risks. These concerns prompted significant amendments to the proposed AI Act agreed and adopted by the European Parliament in June this year.

While the goal of the AI Act is not to hinder innovation in the EU, some of the proposed changes are anticipated to have severe consequences to the region’s fast-growing IT industry. Notably, the European Parliament's decision to deviate partially from the general risk-based approach and impose certain regulatory requirements on any general-purpose AI system, regardless of its intended use, signals a potential departure from the narrow regulation approach.

This revised, technology-neutral definition of an AI system in practice widens the range of regulated technologies. Collectively, these changes suggest the undertone of a broader regulatory framework for AI.

Striking a balance

As the EU navigates this intricate terrain, it finds itself in competition with the UK and the US, both of which tend to adopt a more permissive approach to AI regulation. While the EU's proposals aim to establish comprehensive standards, it must be cautious not to stifle innovation by overburdening the technology with excessive regulations. As the EU seeks to safeguard against potential risks while maintaining a supportive environment for AI advancements, striking the right balance is crucial.

Jeremiasz Kuśmierz is focused on new tech and believes it to be the foundation of future business. His global outlook is highlighted by his participation in multinational deals involving China and Europe. He spreads his work across compliance, employment, dispute resolution, and risk management.