UK Watchdog Warns: AI Models May Quickly Restrict Competition

The U.K. competition regulator has put forth a set of guiding principles to promote healthy competition in the rapidly growing foundation model sector. The Competition and Markets Authority (CMA) released a comprehensive report highlighting the potential for certain companies to leverage their models to command market power and set high service prices. To safeguard consumers and foster robust competition, the CMA has proposed the following principles:

**Accountability**

Developers and deployers of foundation models must be responsible for the outputs they deliver to consumers, ensuring that quality and reliability are maintained.

**Access**

Key inputs should remain readily accessible to all stakeholders, without undue restrictions that would hinder innovation and adoption.

**Diversity**

A dynamic ecosystem of business models is crucial, encompassing both open-source and proprietary systems to cater to different market needs.

**Choice**

Businesses should have ample options to determine how they utilize foundation models, enabling them to select the solutions that best fit their needs.

**Flexibility**

Organizations must retain the flexibility to switch between or adopt multiple models based on evolving demands, encouraging adaptability and resilience.

**Fair Dealing**

Anti-competitive behaviors, such as self-preferencing, tying, or bundling services, should be strictly prohibited to ensure a level playing field for all participants.

**Transparency**

Consumers and businesses should receive clear and comprehensive information about the risks and limitations of AI-generated content, empowering them to make informed choices.

The CMA aims to collaborate with various stakeholders in the AI domain to refine these principles. Key players in this dialogue will include major developers such as OpenAI, Meta, Nvidia, and Google, alongside government representatives, academic experts, and regulatory bodies.

Sarah Cardell, CEO of the CMA, warned that there is a significant risk that AI usage may evolve in ways that compromise consumer trust or result in a market dominated by a few entities with substantial power. “We need to ensure that the full benefits of AI are realized across the economy,” she emphasized.

An update on these principles is anticipated in early 2024. The CMA's exploration of foundation models began in May and aims to conduct an in-depth review of various concerns, including security, safety, and copyright issues.

The CMA’s investigation into AI’s competitive implications comes on the heels of the Sunak government’s strategy to empower regulators to establish rules for AI, promoting a “pro-innovation” regulatory environment.

In a significant report spanning 130 pages, the CMA engaged with approximately 70 stakeholders, including foundation model developers, enterprises utilizing AI technology, and consumer advocacy groups, to shape these initial principles.

Given the rapid pace of AI adoption, the CMA warns that its impact on competition and consumers could manifest swiftly. “Immediate consumer risks include exposure to misinformation, AI-facilitated fraud, or deceptive reviews. In the long run, if a few firms consolidate market power, they may fail to provide the best products and services or impose inflated prices,” the report asserts. “Ensuring that such negative outcomes do not occur is essential.”

Moreover, the CMA acknowledged that effective competition is vital for market health, while underscoring the importance of safeguarding provisions related to data protection and intellectual property rights. This forward-thinking approach aims to foster an environment where innovation thrives while protecting consumer interests.