Control Keys

While generative AI (gen AI) is spurring a quantum leap of innovation in fields from consumer marketing to protein discovery, a rapid consolidation is taking place behind the scenes. And unless policymakers act fast to promote more balance and competitiveness, argues Associate Professor Abhishek Nagaraj, only a few big firms will control the future of the field.

“I have no doubt that the market will be dominated by a few key players,” says Nagaraj, who sounds a warning in a National Bureau of Economic Research working paper. “The big question is how concentrated are we talking about?”

The stakes are high: A landscape controlled by a few AI overlords could lead to less transparency, innovation, and efficiency, stifling potentially more transformative technology. The traditional “moats” that protect startup technology, however, such as patent protections and secrecy around intellectual property, are unlikely to be effective due to the massive edge that large companies already enjoy, Nagaraj says.

In the paper, Nagaraj and colleagues from MIT Sloan and Harvard Business School draw upon the pioneering work of Professor David Teece, who defined two different ways to gain advantage in a competitive environment: appropriability, the ability to guard against copycats, and complementary assets, or control over the ability to transform the innovative know-how into something customers might pay for.

“If I come up with the idea for a drug, for example, I can protect that idea with a strong patent, even if you can already see the idea,” Nagaraj says. With AI technology, the foundational model for gen AI is well understood, making it difficult to protect the core technology. Plus, rapid turnover among Silicon Valley firms makes secrecy impossible. “In California, non-compete clauses are illegal, so it’s quite common for firms to hire from rival companies,” Nagaraj says.

Meanwhile, the complementary assets that large gen AI firms already enjoy are impressive—starting with the massive computing infrastructure needed to run systems, what they call the “compute environment.” Meta alone is acquiring hundreds of thousands of NVIDIA’s state-of-the-art H100 graphics cards at the cost of billions of dollars. “The scale required is mind-boggling,” Nagaraj says.

In addition, big firms are scraping the internet for immense amounts of data to train their models at a level prohibitive for smaller companies. Large players can use the data to set their own performance benchmarks and ethical standards in a way that other companies have little choice but to follow. “These benchmarks are all super subjective and tied to the training data that firms use, so they are implicitly designed in a way that makes the market leaders look good,” Nagaraj says.

Ironically, the gen AI environment has thus far continued to remain competitive due to one of the big players themselves. Meta released an open-source version of its gen AI model Llama, which was accelerated through an accidental leak last year. In its wake, multiple knockoffs, including Berkeley’s Vicuna, Stanford’s Alpaca, and others flooded the market, instantly creating a renaissance in the field. “There are so many ways people are experimenting with it that wouldn’t be possible with just OpenAI,” Nagaraj says. Still, Meta hasn’t been completely open with its training data, and Nagaraj speculates it may try to exert control over the platform through other means, like Google does with Android.

In the meantime, Nagaraj and his co-authors argue for a more hands-on role by policymakers to better control the complementary assets that advantage big companies. One intriguing idea is a public AI infrastructure, similar to the national highway system that aids in interstate commerce. “It could really lower the bar to democratize the compute environment,” Nagaraj says. Regulators could also standardize benchmarks for performance and safety, setting more objective measures that could level the playing field. “We can’t let a small number of companies decide what’s good or what’s safe,” he says.