Artificial Intelligence and competition law

The debate on Artificial Intelligence and the protection of competition is up for consultation in Europe.

Artificial intelligence and competition law

Laura Giménez Rodrigo

Reading time: 4 min

The European Commission launched earlier this year a call for contributions from interested parties on the level of competition in the field of AI  as well as in the Virtual Worlds, which we discussed in this blog a few days ago. Telefónica participated explaining its views on the competitiveness of this sector and anticipating the importance of competition authorities ensuring that all economic operators respect the rules of the game.

Will the transformation that AI will bring be reflected in market competition?

In previous posts, we have talked about the transformation that AI will bring to different areas of our society and economy. It will have a major impact on the daily lives of citizens, businesses and markets in areas as relevant as education, medicine, energy, agriculture, finance, among many others. It may even change our understanding of the world. In fact, this post has been drafted using AI, something that would have been difficult to understand or imagine just a few years ago.

We cannot predict exactly what impact such a revolution will have on the functioning of markets. However, based on past experience in the digital world, it does not seem difficult to foresee that certain players will have the incentives and ability to engage in anti-competitive practices that may harm the market and consumers. Indeed, in order to strike a balance between competitiveness, innovation and responsibility in AI, so that we can all perceive the benefits it will bring, not without risks, it is necessary that all economic actors operate on a level playing field.

 Potential competition risks in the field of AI

Access to certain resources and elements is key to providing services related to Generative AI. Companies with such tools may have high market shares and, because of this position, they may be tempted to impose commercial conditions and/or obligations on their customers that are contrary to competition rules.

For instance, having your own distributed computing infrastructure (cloud) could be considered an essential input in order to build and provide generative AI services. This is also true for the models as a service, which are very difficult to develop as they require an investment of billions to build an LLM (Large Language Model). GPU (Graphics Processing Unit) are also of utmost importance in the provision of AI services since they are cornerstone for training and deploying generative AI models. Furthermore, the data serves as the basis for the development and deployment of generative AI systems and models. This affects their performance, reliability and ethical considerations. Therefore, the quality, quantity and availability or accessibility of data is also a critical factor to be competitive in this sector.

If companies which control all these elements (among some others) refuse to grant such access to third parties (or refuse to provide interoperability) or charge a very high price for them, competition is likely to be weakened.

Looking at previous European Commission decisions on abuse of a dominant position in the digital sector, it is not difficult to conclude that AI could be another example where dominant companies may seek to favour their own products and services to the detriment of those of smaller economic operators and, ultimately, to the detriment of consumers and competition (the so-called “self-preferencing” practices).

As mentioned above, such companies might leverage their dominant position from one market to an adjacent one (upstream or downstream). They could impose technical conditions for the integration of third parties’ services with their own, make payments conditional on their customers installing or using only the services of such dominant undertakings, make the sale of one product or service conditional on the purchase of another, or force or induce their customers to purchase two or more products together.

Competition authorities should therefore monitor that such breaches do not take place. 

Compliance with competition rules is essential to achieve benefits for all

Despite the risks associated with the development of generative AI, there are many benefits to be gained. The difficult challenge will be to ensure that technological advances are made in compliance with competition law. Only then will society benefit from a greater variety of innovative products and services, at better prices and of higher quality.

As with any change, markets will have to adapt to the disruption that AI will bring. Society and citizens will have to do the same in the various activities in which AI plays a role. We may do things differently or use AI to be more efficient – Will markets do the same without engaging in anti-competitive practices? Some believe that certain aspects of life will be beyond the reach of AI.  We hope that the reader’s ability to intuit or discover whether this post has been written with AI, as we pointed out at the beginning of the text, is one of them.


Communication

Contact our communication department or requests additional material.

Telefónica Centenary logo Celebrate with us the Telefónica Centenary
START THE ADVENTURE