In brief
The Canadian Competition Bureau (“Bureau“) has announced a significant initiative to explore the impacts of artificial intelligence (“AI“) on competition in Canada. On 20 March 2024, the Bureau published a discussion paper, entitled “Artificial intelligence and competition” (“Discussion Paper“), and has requested feedback from the public. The Bureau is seeking to better understand how AI can affect competition in Canada. It is also seeking to determine how it can better prepare to address potential competitive harm from AI, and also promote competition in AI markets. The outcome of the consultation may have significant implications for both businesses that develop and those that use AI technologies. Submissions are due by 4 May 2024.
Key takeaways
AI is affecting many sectors of the economy and is increasingly being embraced by Canadian businesses and Canadian consumers alike.
Recognizing that it is imperative to comprehend and address the implications of AI, the Bureau issued the Discussion Paper with the aim of fostering thoughtful and informed national dialogue in order to help the Bureau deepen its understanding of how competition is developing in AI markets, and how the Bureau can enforce and promote competition in these markets.
The Discussion Paper provides an overarching definition for the term “AI” and explains the different technologies that make up AI (Section 1). It also describes the various markets that are involved in the creation of an AI product or service (Section 2), including markets for AI infrastructure, markets for AI development, and markets for AI deployment.
The Discussion Paper then explores several considerations for how AI can affect competition, particularly in the context of the Bureau’s enforcement areas, such as (i) mergers and monopolistic practices, (ii) cartels, and (iii) deceptive marketing practices (Section 3).
Regarding mergers and monopolistic practices, the Bureau’s mandate includes investigating and acting when it discovers abuses of market power or mergers that are likely to substantially lessen or prevent competition. Certain characteristics of AI markets can engage these concerns. A few of the issues identified in the Discussion Paper are noted below:
- Barriers to Entry. The Bureau observed that participation in AI development markets heavily relies on the ability of parties to access data and “compute inputs” (e.g., AI chips and cloud services) and, as a result, difficulties in accessing these necessary inputs can serve as a barrier to entry for new firms. While public data is currently highly prevalent, the Bureau noted that the concern will become especially pronounced if and when proprietary data becomes a necessary input in order to participate in AI markets. To this end, the Bureau remarked that large technology firms already have access to substantial quantities of proprietary data that is suitable for AI development (which has been gathered from business activities in other markets that they take part in), that may provide them with a competitive advantage.
- The Bureau further noted that new entrants may face other barriers to entry, including the high concentration of several dominant technology firms in the cloud computing and AI chip markets; high upfront costs of compute inputs; and the current shortage of graphics processing units (GPUs) used for AI.
- Economies of Scope and Scale. The Bureau also observed that there are substantial upfront costs associated with AI development markets and that large technology firms may have a competitive advantage in achieving economies of scale due to existing in-house resources (including human expertise and computing). It noted that this may create difficulties for new market participants (many of whom do not have similar resources) from entering the market or competing effectively.
- Predatory, Exclusionary, and Discriminatory Conduct. The Bureau also expressed concern that parties could use AI to engage in predatory, exclusionary, or discriminatory conduct. Predatory conduct can occur when a firm incurs short-term losses to eliminate a competitor and gain market power. By using AI, for example, the Bureau speculated that a firm could engage in such conduct by using the technology to identify “at-risk” customers who may switch and provide them with targeted below-cost price cuts. AI markets could also be used to engage in exclusionary conduct, by preventing another firm from operating in the market, particularly due to the importance of vertical relationships in such markets. The Bureau noted that vertically integrated firms could supply an important input for participation in AI markets (e.g., data, compute, and AI models), while also competing in the same downstream markets as the firms that they supply, which can lead to exclusionary conduct concerns. AI could also autonomously engage in discriminatory conduct, which the Bureau noted highlights the need to ensure proper human oversight.
- Merger Considerations. The Bureau further observed that certain characteristics of AI markets can create new means for firms to obtain market power through mergers. As such, the Bureau noted that mergers in the AI space, of any form, that involve a firm that supplies compute inputs (such as AI chips and cloud services) could warrant additional scrutiny due to the existing high levels of concentration in these markets. The Bureau suggested that mergers in AI markets could also require additional scrutiny given that large established firms may seek to acquire emerging competitors as a means of preventing or lessening competition.
Regarding cartels, which are formed when two or more firms agree to fix prices, allocate markets, or restrict supply, the Bureau highlighted several concerns that arise in AI markets, including the following:
- The Facilitation of Cartel Behaviour. The Bureau identified the risk that AI may be used to implement or sustain cartel agreements. For example, as explored in the Discussion Paper, firms could use AI to automatically implement agreed-upon prices, or communicate with cartel members, or punish members who deviate from a cartel agreement. AI could also be used to increase market transparency and increase the frequency of interactions between competitors, which the Bureau indicated are two market conditions commonly associated with a greater ability and incentive for cartels to form and operate. The Bureau further noted that if multiple competitors purchase and use the same AI technology from a third-party supplier, “hub-and-spoke” arrangements can form where a supplier of AI can leverage the technology in order to facilitate a cartel among competitors it supplies to. The Bureau observed that there is also increased research and discussion around the possibility of AI to generate “tacit algorithmic collusion”, in which AI that is used by horizontal competitors may autonomously learn and implement coordinated strategies, similar to what may be implemented by an explicit cartel.
Regarding deceptive marketing practices, the Bureau highlighted the importance of consumers receiving truthful information in order to help them make informed buying decisions, and certain issues that may arise with AI, including:
- The Scaling Up of Deceptive Marketing Practices. The Bureau expressed concern that the automated nature of AI, and recent developments in generative AI, could potentially be exploited by parties to engage in certain types of deceptive marketing conduct. For example, AI can automate and scale certain conduct, such as deceptive telemarketing or prize notifications, or make deceptive marketing more “convincing” (e.g., through “deepfake” (digitally altered) voices, images, or videos). To this end, in March 2024, the Bureau addressed the rise of fraud and the use of AI in order to help parties make their frauds more convincing, including through the use of chatbots that spread fake reviews and voice cloning.
In conclusion, the Bureau is seeking feedback on the above issues and on a number of subtopics in each of these areas, including any additional technologies that the Bureau should consider in order to advance its understanding of AI, the characteristics of AI markets in Canada, and whether there are any other factors that can affect competition in the AI sector.
