The top U.S. antitrust regulators, Federal Trade Commission (FTC) Chair Lina Khan and the Department of Justice’s antitrust chief Jonathan Kanter, recently argued that government action may be warranted to prevent large technology companies from using anticompetitive tactics to protect their standing in the emerging artificial intelligence (AI) market. The FTC later followed these comments with an article outlining its concerns and the potential avenues for anti-competitive behavior. However, in advocating for antitrust action, both regulators overlook the recent successes of emerging AI companies; fail to acknowledge large tech companies’ significant contributions to AI research and development, which foster competition; unjustly criticize legitimate business practices; and discount the benefits of economies of scale.
First, when Kanter claims that the AI industry has a “greater risk of having deep moats and barriers to entry,” he fails to consider that new companies, such as OpenAI and Midjourney, started the current generative AI frenzy, ahead of their more established peers. These and other successes show that companies of all sizes can enter the market with new and innovative products without facing undue barriers. Furthermore, the continued development of new AI products suggests that established companies’ competitive advantages cannot keep new entrants out of the market. Indeed, while the FTC argues that the “high cost of entry to creating a pre-trained base model may lead to a market where the highest quality pre-trained models are controlled by a small number of incumbents,” they later note that many of the recent open-source image generation models have become better than the models that inspired their creation. Moreover, any company can scrape the web for data to train its AI systems, so large tech companies do not have a monopoly on data.
Second, while Chair Khan worries about larger companies “becoming bigger and really squelching rivals,” such companies have not undertaken any unreasonable or unfair practices that would justify government intervention. Many large companies, such as Google and Meta, have since released their own generative AI systems instead of resorting to anticompetitive behavior or pursuing emerging companies. In fact, both companies have open-sourced some of their systems and provide extensive research and datasets for other parties to use. These contributions of research, code, and datasets make it possible for potential rivals to build their own competing AI services.
Third, though the FTC argues that established companies have the resources and potential to undermine new companies, the examples the FTC gives constitute legitimate business practices— not anticompetitive behavior. In its article, the FTC lists established companies’ ability to hire better talent as a competition concern as well as the fact that older companies have had more time to hone their data collection skills. The FTC also suggests that companies with both computing services and generative AI systems may give “discriminatory treatment to themselves and their partners over new entrants.” All three of these factors, however, are common business practices. Established companies have the resources to hire talented workers and, by the nature of their established status, have had more time than new companies to develop data collection methods. Moreover, established companies acting in their own self-interest with the provision of their goods and services does not automatically mean they are partaking in anti-competitive behavior.
Fourth, the FTC warns incumbent firms against bundling “new generative AI applications with existing core products.” While bundling and tying can be anticompetitive in some circumstances, in many other cases it improves consumer welfare. Virtually all tech companies are moving quickly to integrate AI into their products because this benefits consumers, such as with better search, integrated writing tools in word processing applications, and automatic image generation in presentation software. A blanket prohibition on incumbent companies integrating AI into these tools would significantly forestall the adoption of AI in the United States, leaving consumers without access to the benefits of AI and making U.S. businesses less competitive with their rivals abroad.
Finally, in targeting large companies operating in good faith to prevent market concentration, the FTC and DOJ discount the benefits of economies of scale. Large companies can operate more efficiently than smaller firms and offer lower prices than their smaller peers. Indeed, in a 2021 study, the Organization for Economic Cooperation and Development (OECD) found that a 10 percent increase in market concentration is associated with a 2.3 percent decrease in industry prices. While the FTC and DOJ should take action against any company employing anti-competitive methods, unnecessarily constraining large companies prevents consumers from obtaining these lower prices.
The AI industry is emerging and evolving rapidly, and neither the FTC nor the DOJ has been able to point to any evidence of anticompetitive behavior. Rather than invent or speculate about problems that do not exist, competition regulations should allow the AI industry to continue to develop new and innovative products without unwarranted restrictions so that both U.S. businesses and consumers can access the benefits of AI.