People can’t stop talking about AI and the various ways it is impacting our lives. Discussions about its potential and risks dominate all parts of our culture and society. It’s almost impossible to read a newsletter or attend a conference without the technology earning at least a mention, if not several headlines or keynote speeches.
The updates to AI tools, and especially to Large Language Models (LLMs), are happening so quickly that only the reporters and tech enthusiasts following them closely are able to keep up—a sign of a hyper-competitive market for new AI products. Another is the vast number of highly-valued businesses in AI. At the end of 2023, startup news outlet Failory identified 126 “unicorns” in AI, which is a startup with over $1 billion valuation.
Yet over-zealous antitrust enforcers in Washington, DC, are raising competition concerns, and the Federal Trade Commission (FTC) even launched a full inquiry into the tech AI industry over these allegations in January. Such a preemptive action taken against a competitive industry at the expense of other important duties for the FTC (such as clamping down on an epidemic of robocalls, fraud reports, etc.) is indicative of their progressive antitrust shift. Rather than focus on protecting consumers from actual, observable harm that they are experiencing, Chair Lina Khan’s FTC and Jonathan Kanter’s antitrust division at the Department of Justice (DOJ) have preferred to attack American businesses even if it comes at the expense of the consumer.
The FTC’s inquiry is based on a concern that large tech companies could one day dominate the AI market, or at least LLMs. However, the vast amount of innovation and development by new companies in this space proves otherwise. Of course, as technology moves from groundbreaking new science to a product or service like any other, consolidation can occur and indeed be beneficial for consumers. Only when important markets get particularly concentrated or uncompetitive does the government traditionally intervene.
But under the progressive antitrust string pulling, government intervention is a weapon to be wielded in all markets at any time, regardless of how competitive they are.
At the end of their inquiry, the FTC could end up accusing leading American AI developers of harming competition in the market and take punitive action—at a severe cost to U.S. leadership in this new tech.
Yet, if we look at the evidence, including actual dynamics from the AI market, we quickly see that the FTC’s competition concerns are at best misplaced, if not an uncharitable excuse to once again attack American enterprise.
The market for LLMs illustrates a starkly different narrative from the monopolistic premonitions painted by advocates of intervention.
Firstly, the landscape of AI firms, especially those focusing on LLMs, is far from sparse. Recent funding rounds alone prove the competitiveness of the market. OpenAI’s huge $10 billion investment from Microsoft underscores the fact that many leading tech companies and their internal AI projects have fallen behind new competitors in developing artificial intelligence. Similarly, Cohere’s $270 million and Anthropic’s $450 million Series C rounds, and the global spread of AI investments in Tel Aviv’s AI21 Labs to Germany’s Aleph Alpha, all show how competitive and diverse the field for developing AI technology is all over the world.
The surge in success of open-source LLMs also introduces and maintains a competitive check to the AI market by ensuring high-quality software is usable by companies who are yet to catch up to the industry’s current leaders. By lowering entry barriers, fostering innovation through collaboration and setting industry benchmarks, open-source models ensure a competitive environment that encourages continuous improvement and diversification of products. With open-source models at their dispense, newcomers can more easily challenge incumbents, thereby preventing monopolistic stagnation.
So, even if the AI industry were to become more concentrated, open-source models mean that if the businesses in that industry stop catering to consumer needs, the market will be inviting for new competitors.
Open-source algorithms are offered by well-known tech businesses like Meta and their LLaMa models, but also by less widely well-known organizations like the Technology Innovation Institute. The performance of their most recent model, Falcon180B, performs “just behind” ChatGPT’s market-leading GPT 4.
Research and development (R&D) can be a key metric to determine whether an industry is competitive or not. Uncompetitive industries lack the market incentives for businesses to spend on risky R&D projects, and as a result, the rate of technological advancement in the industry slows to the detriment of consumers. If only a fraction of recent investments in AI startups is spent on R&D, that still represents a huge spend on innovation in the industry. In Q3 of 2023, Crunchbase added a whopping 15 new AI companies to their “Unicorn Board” where they record all the tech startups that have reached a valuation of at least $1 billion.
The AI market even has competition from universities, with Stanford University’s Alpaca proving that success in this market is enabled by strong research just as much as it is by big budgets. These efforts signify a collective push towards safer, more innovative and accessible AI technologies, driven by the competitive urge to lead in the next generation of tech advancements.
Of course, we know that overzealous government intervention ends up harming consumers. A recent report by the Buckeye Institute showed how federal and state efforts, such as President Biden’s executive order on AI, have created a climate of uncertainty and increased costs, particularly for smaller AI firms, stifling competition and innovation. This heavy-handed approach on AI’s application in healthcare not only hampers the development of life-saving technologies but also creates a regulatory landscape that is more favorable to large, established firms, leading to a less competitive market. As in this situation, overzealous antitrust enforcement will likely end up hampering the FTC’s given goal of protecting consumers.
We shouldn’t be surprised that Lina Khan’s FTC has preemptively involved itself in a competitive economy in search of prominence and relevance. As the agency’s radical leadership struggles to identify what area of the American economy they don’t need to intervene in, what may surprise us are the facts: how competitive the AI marketplace truly is.
Image generated by NetChoice using ChatGPT’s DALL-E.