Clifford Chance

On 6 June 2025, the Japan Fair Trade Commission ("JFTC") released an interim report on generative AI, identifying problematic practices and potential antitrust issues such as access restrictions and tying.

Background

In October 2024, the JFTC published a discussion paper titled 'Generative AI and Competition' (the "Discussion Paper") to solicit information and opinions from a wide range of interested parties pertaining to generative AI and competition. The discussion paper examined the structure of the generative AI sector, categorising it into three layers: the infrastructure layer, the model layer, and the application layer.

The Discussion Paper raised five key areas where the JFTC thinks there may be issues:

1. Access restrictions and exclusion of competitors
2. Self-preferencing
3. Tying
4. Parallel conduct (by using the same generative AI)
5. Acquisition of expert talent, based on the concern that a dominant company may try to acquire many highly skilled professionals, thereby creating an entry barrier in the generative AI market.

After the publishment of the Discussion Paper, the JFTC analysed trends based on the information and opinions collected through public comments and interviews with businesses and experts, compiling these insights into this interim report (the "Report"), which was published by the JFTC on 6 June 2025.

Potential Antitrust Issues Considered in the Report

The Report delves into two issues: (i) access restrictions and exclusion of competitors, and (ii) tying, as these actions could exclude new entrants and existing competitors, or diminish their business opportunities, leading to market foreclosure.

• Access Restrictions: The JFTC has identified potential issues when dominant companies in computing resources (such as GPUs), data, or expertise, restrict access to these resources, resulting in competitors or potential new entrants being foreclosed. Although such conduct has not been found, concerns persist about the challenges in securing alternative suppliers in the future.
• Tying: The JFTC also highlights that potential issues may arise if generative AI model providers who have a strong position in a specific digital service market integrate its generative AI model into their digital service and offer it to users as a new digital service.

The JFTC further addresses other issues raised in the Discussion Paper, namely: (i) self-preferencing, where companies providing generative AI model might develop them to favour their own goods and services in inference results, (ii) parallel conduct using generative AI, where the shared use of underlying data and algorithms could result in similar pricing strategies and production targets; and (iii) acquisition of highly skilled professionals. However, the JFTC received no comments indicating problematic practices in relation to these issues.

The JFTC has indicated that it will continue its market study on generative AI, and will update and supplement this report as it gains a better understanding of the actual market conditions.

Practical Implications

This Report is in line with the approach taken by other competition authorities in assessing the potential impact of generative AI on competition, such as the European Commission’s policy brief published in September 2024 and the UK Competition and Markets Authority’s updated paper on generative AI, published in April 2024.

As the competitive landscape in generative AI continues to evolve rapidly, and as these technologies become increasingly important to a wide range of business activities, companies operating in or relying on generative AI should closely monitor developments in antitrust enforcement and policy in this area. Competition authorities, including the JFTC, are likely to refine their approaches in response to market changes and emerging business practices.