Initial review of generative AI and competition law: Insights from European competition authorities

Recently, the intersection of generative artificial intelligence (AI) and competition law has become a focal point for competition authorities worldwide. With the imminent adoption of the EU AI Act in the European Union, the spotlight on AI-related regulatory issues has intensified. Notably, European competition authorities have taken significant strides to comprehend the competitive landscape in AI-affected markets. Their efforts aim to address potential harms and pave the way for future legislation, investigations, and remedies.

Below, we have collected some of the key examples in order to provide market players and stakeholders with a panoramic overview of the possibilities and risks they represent.

Recent Initiatives by European Competition Authorities

1. European Union

The European Commission[1], recognizing the transformative impact of AI, has launched a call for contributions. Market participants are invited to share their views voluntarily. The deadline for submissions is 11 March 2024. This initiative seeks to gather insights from industry experts, academia, and consumer organizations regarding competition aspects in the AI domain.

2. United Kingdom

The UK Competition and Markets Authority (CMA)[2] conducted an initial review specifically focused on AI Foundation Models. Their findings were published in a comprehensive report in September 2023. The report sheds light on the competitive dynamics surrounding AI models and their implications. Furthermore, the CMA plans to update this review by March 2024, reflecting the dynamic nature of the AI landscape.

3. France

The French Autorité de la Concurrence (Autorité)[3] has started inquiries ex officio and launched a public consultation on the generative artificial intelligence (AI) sector. The Autorité will examine the competitive functioning of the sector, especially the practices of major digital players that already control key inputs or adjacent markets, such as cloud services, data and skilled workforces, and their investments in innovative start-ups specialised in generative AI. Stakeholders are invited to respond to the Autorité’s questions and submit their answers by 22 March 2024.

4. Portugal

The Portuguese Audoridade da Concorrência (AdC) [4] has published an issues paper about the role of competition and generative AI in November 2023. In that paper, the AdC noted that the focus of competition in the context of generative AI will necessarily have to be on the following cornerstones: access to data, access to cloud computing or specialized hardware and access to base models.

5. Hungary

The Hungarian Gazdasági Versenyhivatal (GVH)[5] has embarked on a market study to delve into AI markets from both antitrust and consumer protection perspectives. In Q1 2024, the GVH will distribute questionnaires to various market players. Additionally, voluntary contributions from stakeholders are encouraged. This study aims to unravel critical aspects such as data importance, entry barriers, interoperability, and the role of open-source models.

The Role of Companies in Shaping AI Antitrust Policies

As companies actively engage in AI development, deployment, and application, their participation in the ongoing discourse about AI’s antitrust implications becomes crucial. As per the initial studies and reviews cited above the key issues include:

• Data Dominance: The availability and control of data play a pivotal role in AI markets. Understanding data dynamics is essential for assessing market power and potential anticompetitive behaviour.
• Barriers to Entry: Identifying barriers that hinder new entrants from competing effectively in AI markets is vital. Addressing these barriers fosters healthy competition and innovation.
• Interoperability: Ensuring interoperability among AI systems promotes fair competition. Standards and protocols that facilitate seamless integration are critical.
• Open Source Models: The role of open-source AI models in fostering competition deserves attention. Balancing proprietary and open-source approaches is essential for a vibrant AI ecosystem.
• Policy: the more general question of the competition policy approach is certainly the “toughest nut to crack”: namely whether to take a hands off approach and leave this new and innovative industry to the market forces or rather to take a closer scrutiny already now in order to avoid inevitable future concentration (just as it happened with digital markets in the 2010s).

Conclusions

In conclusion, as the AI landscape evolves, collaboration between competition authorities, industry players, and other stakeholders is paramount. By actively participating in discussions and contributing valuable insights, companies, academic institutions and other market stakeholders can shape antitrust policies that foster innovation, fairness, and consumer welfare in the AI era. Based on these inputs, the authorities can then make further decisions on how actively to tackle this very innovative and exciting market segment.