On July 23, the European Commission, US Department of Justice, US Federal Trade Commission, and UK Competition and Markets Authority released a joint statement setting out shared principles on competition in markets that deploy generative AI foundation models and other AI products.
Although the joint statement recognizes the immense potential for AI to benefit individuals, innovation, and the economy, it also enumerates risks that the authorities believe that further deployment of AI could pose to competition and consumers.
The statement lists primary potential risks to competition from generative AI: (i) concentrated control of key technology inputs for developing foundation models (specialized chips, substantial compute, data at scale, and specialist technical expertise); (ii) entrenched or extended power in AI-related markets where large incumbent digital firms already enjoy advantages; and (iii) incumbent firms using partnerships, investments, and other relationships related to developing generative AI to undermine competitive threats and steer market outcomes in their favor.
The statement also posits other ways in which AI could harm competition, particularly based on algorithmic pricing models, and states that the authorities will be monitoring and addressing risks that may arise in other AI contexts, beyond generative AI.
The statement includes an interesting discussion of the authorities’ principles for protecting competition in the AI ecosystem. The authorities suggest that they will seek to ensure “fair dealing” among market participants; interoperability of AI products and services; and that businesses and consumers are able to choose from diverse products and business models. Notably, the statement does not explain how these objectives will fit with the domestic antitrust laws in each jurisdiction. Not all of the jurisdictions, for example, recognize general “fair dealing” or “interoperability” obligations. Although exclusionary conduct by firms with substantial market power can violate the antitrust always in each of the jurisdictions, there is no general principle that dominant firms must engage in “fair dealing.” The EU recently introduced the notion of “fairness” in the digital ecosystem under its laws, but it is debatable whether this aspect of the Digital Markets Act is truly an antitrust concept.
Related to consumer protection, the statement notes that the agencies will monitor threats that may derive from the deployment of AI, particularly where firms deceptively use consumer or business customers’ data to train their models. This reinforces the growing importance of the principle of personal data protection with existing legislation like the EU’s GDPR and recently passed AI Act. Furthermore, in the joint statement, the agencies note the importance of informing consumers about when and how an AI application is employed in the products and services they purchase or use.
