Competition Concerns in the Age of AI

Updates

中文版

June 12, 2023

Competition Concerns in the Age of AI

Across every industry, companies are leveraging machine learning to derive valuable insights without extensive employee involvement. These groundbreaking capabilities are creating an upheaval in how companies engage with competitors and consumers.

Adept competition and consumer protection attorneys help companies capitalize on the opportunities artificial intelligence (AI) presents while navigating the terra nova of regulatory and litigation risk. Although it is incorrect to approach AI as a black box, the complexity of AI systems can make reasoning opaque. This means linkages between AI outputs and rational business justifications risk being obscured or even lost entirely.

Yet regulators are unlikely to excuse consumer and competitive concerns merely because an organization cannot explain why certain actions were taken and others were not. Legal exposure exists under the Sherman Antitrust Act, Federal Trade Commission (FTC) Act, and Robinson–Patman Act, as well as state antitrust and consumer protection laws. By implementing policies and processes that preserve human control and accountability, organizations can minimize legal exposure and avoid unintended consequences.

A proactive and customized approach is critical. AI affects competition and consumers in countless ways, including when used for core business functions.

Pricing

AI helps companies make pricing decisions by responding quickly to instantaneous changes in demand, inventory, and input costs. By synthesizing and summarizing vast amounts of complex data, AI can significantly aid in building and adapting pricing policies. But the outcomes that AI-assisted pricing generates can also be seen as facilitating per se unlawful collusion, such as price fixing or bid-rigging. According to FTC Chair Lina Khan, AI "can facilitate collusive behavior that unfairly inflates prices."

These concerns may arise directly or indirectly from using AI to perform a diverse array of activities, such as benchmarking, disaggregating information, signaling, exchanging information, or analyzing pricing trends. Pricing algorithms, for example, may raise antitrust issues when competitors use them to enforce an advance agreement, algorithm vendors initiate or organize an agreement, companies apply algorithms to dramatically raise prices, or even when competitors independently employ algorithms that subsequently engage in collusive conduct. The U.S. Department of Justice's Antitrust Division highlights that "the rise of data aggregation, machine learning, and pricing algorithms ... can increase the competitive value of historical data" and warrants "revisiting how we think about the exchange of competitively-sensitive information."

Purchasing

Deploying AI to determine what and how much to buy and from whom to purchase presents attractive opportunities for streamlining and optimizing operations. On the other hand, AI-driven purchasing decisions that lack appropriate guardrails can create an undue inference of price discrimination, collusion, or concerted exclusionary conduct. Foreclosure or denigration of competitors, for example, can create or enhance market power. In the current enforcement climate, merely obtaining a better purchase price than a competitor is not without risk. The same is true for terminating a supplier or entering exclusive contracts. Contemporaneously documenting the procompetitive rationale for these practices is an important tool to minimize potential exposure. This requires effective and ongoing training with human intervention.

Production

AI is an important asset when making production and output decisions. But without effective monitoring, AI-driven recommendations across firms can create an undue inference of unlawful agreements to curtail production or allocate markets, products, or customers with competitors. Such results are likely to attract the attention of antitrust regulators and private litigants. Companies should therefore exercise appropriate supervision of AI tools when making decisions related to output and production volume. This includes regularly tracking the impact of implementing AI recommendations on company and industry production levels.

Distribution

Reevaluating distribution partners is an ongoing exercise for many businesses. AI can make sense of a vast amount of partner performance and product investment data to help companies decide which distributors and resellers to prioritize and reward and which to de-emphasize. But preferencing certain partners can raise antitrust issues related to the exclusion or foreclosure of competitors, products, or markets. Antitrust guidance and training are crucial to avoid exposure under theories of refusals to deal, exclusive dealing, conditional dealing, tying, and multiproduct bundling. Regulators are closely scrutinizing parallel exclusionary conduct that may cause aggregate harm, as well as collusive conduct involving lockstep changes to distribution practices.

Capital Expenditures and R&D

AI is an important tool for helping companies decide when and where to invest in capital expenditures and how to allocate their R&D budget. Antitrust risk should be considered when leveraging AI for building market or segment strategies, generating new product or market ideas, or dissecting profitability by sector or customer compared to the competition. Using AI to inform such decisions without appropriate guardrails can lead to an unintended unreasonable restraint of competition.

Employees and Contractors

Conduct affecting workers is at the forefront of the enforcement agenda. Using AI to digest data on employee compensation across an industry or geographic area can result in a salary structure that appears collusive. Without ongoing human oversight, AI-assisted analysis of another company's employee compensation data can lead to parallel conduct and lengthy antitrust litigation. This holds true regardless of whether those companies compete to sell their goods or services to customers. A company's competitors in the labor market are generally much broader than its competitors for products and services.

Consumers

Fraud is a key concern surrounding generative AI, which refers to tools used to craft apparently authentic—but ultimately false—text, images, music, videos, and voices. Companies that run advertisements as part of their marketing strategy must be vigilant in ensuring that AI tools are not misleading or confusing consumers. The FTC warns that businesses must implement "durable, built-in" precautions to prevent fraud and deception through doppelgängers, deepfakes, and chatbots. Those that fail to do so risk lengthy enforcement actions.

Experienced competition, consumer protection, and technology lawyers can empower companies to thrive in the age of AI through five core services:

Advise on processes to identify and preserve human business justifications and accountability when utilizing AI tools.
Build policies on the use of AI to ensure data integrity, competitive vigor, accuracy, transparency, and ongoing monitoring.
Create antitrust guardrails tailored to the use and proliferation of AI tools.
Determine competitive implications of using certain data, models, or AI tools for pricing, production, purchasing, and product allocation decisions.
Evaluate whether AI tools can be used for their desired function based on market and competitive dynamics.

A previous version of this Update appeared in TechCrunch+.