OECD paper examines competition effects of AI adoption in downstream markets
AI may support competition by reducing costs and enabling innovation, but effects depend on data, compute, and firm capabilities.
The OECD has published a competition policy paper examining how AI adoption, including generative and agentic systems, may affect competition in downstream markets.
The paper focuses on how firms use AI as an input into production, service delivery, logistics, and customer engagement, rather than on competition in AI infrastructure or foundation model development. The paper states that AI may support competition by lowering barriers to entry, reducing minimum efficient scale, and supporting product differentiation and innovation.
Generative AI can automate or accelerate cognitive tasks such as writing, coding, summarisation, translation, planning, image generation, and customer support. According to the paper, these tools may allow smaller firms and start-ups to operate with lower staffing and operational costs.
The paper identifies potential gains from AI-enabled personalisation, predictive analytics, and cost reduction. AI tools can help firms offer tailored services, reduce operating costs, and improve matching between consumers and suppliers.
The OECD said the effects of AI adoption may vary depending on factors such as firm size, sector exposure, access to data, and computing resources. Adoption costs, integration challenges, access to data and compute, firm size, and sector exposure can all shape whether AI strengthens competition or reinforces existing market advantages.
The paper identifies competition concerns, including algorithmic collusion, personalised pricing, bundling, and dependence on large model providers or cloud platforms. It also warns that dependence on a small number of model providers, cloud platforms, or proprietary data sources could limit downstream contestability.
The paper describes agentic AI as an emerging issue for competition authorities. Systems made up of multiple coordinated AI agents could reshape search, workflow automation, customer engagement, and consumer choice, while raising new questions about liability, auditability, oversight, and market structure.
The OECD said competition authorities may require a combination of enforcement, market monitoring, regulation, and cooperation to address AI-related market developments. It also identifies areas for further research, including sector-specific impacts in health, finance, professional services, platform services, search, logistics, and creative industries.
Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!