Earlier this year, the European Union agreed on new, stringent rules to regulate artificial intelligence. The landmark legislation, called the EU AI Act, aims to place Europe at the forefront of the global AI race while ensuring safeguards for humanity’s interests.
The Act introduces a structured approach to AI governance, by categorising AI applications into four risk profiles – and imposing corresponding regulations. Non-compliant companies face hefty fines: €35 million or 7% of global revenue.
The implications of this Act are profound, not only for the technology sector, but also for the broader political and economic landscape of Europe. So we caught up with Philip Meissner, Chair of Strategic Management and Decision Making at ESCP in Berlin, to provide a critical perspective on the impact of the Act.
Not having the latest AI will be equivalent to not being connected to the energy grid.
Key provisions of the AI Act
The Act categorises AI applications into four risk profiles:
- Unacceptable risk: AI technologies that pose an unacceptable risk, such as social scoring and behavioural manipulation, are outright banned.
- High-risk applications: This category includes AI used in education, transport and employment. These applications face tough regulations to ensure safety.
- Limited risk: AI applications with limited risk are subject to lighter regulations. For example, someone using a chatbot needs to be fully aware they are ‘talking’ to a machine.
- Low risk: These applications, such as spam filters and video games, face minimal or no regulatory constraints.
Additionally, the Act covers general-purpose AI, such as ChatGPT, by proposing transparency criteria for these models and ensuring compliance with copyright law. And, crucially, the regulation applies to any AI affecting individuals in the EU, regardless of where the developing company is based.
Impact on AI development and deployment
Professor Meissner, the Founder and Director of the European Center for Digital Competitiveness, acknowledges the argument made by proponents that the regulation aims to foster trustworthy AI within the EU. However, he also expresses concerns about the potential negative impacts on AI deployment and the EU’s competitiveness in this vital technology sector.
“Companies from outside the EU may decide to no longer roll out their technologies in Europe due to the regulation. We already see today that Meta’s Llama 3 model is not available in the EU,” Meissner notes, citing the AI assistant developed by the American tech giant.
He adds that this reluctance could lead to a scenario where the latest AI technologies are unavailable in the EU, hampering the competitiveness of European individuals and businesses. Meissner stresses that “not having the latest AI will be equivalent to not being connected to the energy grid”, highlighting the crucial role AI plays in modern technological infrastructure.
Furthermore, Meissner warns that the Act could exacerbate the divide in AI investments between the EU and the rest of the world. For example: from 2018 to Q3 2023, EU AI companies received nearly €32.5 billion in investments, while US AI companies garnered more than €120 billion. Stricter regulations might continue this trend, Meissner says, leading companies to invest in regions with lighter regulatory environments.
Priorities for the next European Commission
As the dust settles on the recent European Parliament elections – in which citizens in all 27 EU member states cast their ballots in early June – the priorities for the next European Commission are coming into sharp focus.
Meissner argues for a recalibration of economic policy, underscoring the need for less regulation and more freedom to spur growth and innovation. He outlines three key steps:
- Deregulation: Meissner advocates for reducing existing regulations and ensuring new rules, like the AI Act, are not overly burdensome. “The new Commission should focus on reducing existing rules and interpreting the newly-created ones in a way that does not turn them into a bureaucratic monster for the economy,” he suggests.
- Competition policy: To enhance the competitiveness of European companies, some of which fear falling behind their US peers, Meissner calls for a shift in EU competition policy to allow European firms to grow and compete globally. He points out that the largest European telecommunications companies are significantly smaller than their American counterparts, limiting their ability to invest in future technologies.
- Commissioners for ‘future technology’: And lastly, Meissner highlights the need for a dedicated focus on “future technology” within the Commission. He suggests appointing commissioners for AI, technology, or competitiveness, which would signal a commitment to innovation and technological advancement.
“The new Commission should focus on reducing existing rules and interpreting the newly-created ones in a way that does not turn them into a bureaucratic monster for the economy.”
Balancing innovation and regulation
Ensuring the EU remains a global leader in AI while balancing innovation and regulation is, overall, a complex challenge. Meissner stresses the importance of a unified regulatory framework between the EU and the US to prevent regulatory arbitrage, where companies relocate to regions with more favourable rules.
He acknowledges that the EU is currently a leader in AI regulation, not the technology itself (the US and China are further ahead in that respect), and argues for a more open regulatory environment to support the competitiveness of AI companies in Europe.
“One way to enable a more open regulatory environment would be to not enforce the AI Act as strictly, or be less strict when it comes to assessing applications as high risk versus limited risk,” Meissner suggests. This approach would help retain leading AI companies within the EU, without putting them at a disadvantage compared to global rivals.
The EU’s Digital Decade initiative
The AI Act is, meanwhile, aligned with the EU’s Digital Decade policy programme, which aims to drive digital transformation and innovation while ensuring technology benefits society and the economy. However, Meissner believes that the EU’s focus on regulation has overshadowed the importance of economic and technological competitiveness.
He calls for a shift towards a more business-friendly policy to achieve technological sovereignty and economic growth. “This may be the last chance to thrive towards sovereignty when it comes to future technologies and the wealth creation they enable,” Meissner warns. Failing to seize this opportunity could leave the EU dependent on external technologies for decades to come.
Ultimately, balancing regulation with the need for innovation and competitiveness will be key to ensuring the EU remains a leader in the global AI race.