A computer science expert at UW-Madison expects major AI firms will face a dwindling economic advantage as widely available models become more powerful, while companies running data centers will thrive.
Patrick McDaniel, a professor in the university’s School of Computer, Data and Information Sciences, was a featured speaker during the latest UW Now livestream held by the Wisconsin Alumni Association.
“The people that are building models for AI aren’t necessarily the companies that are going to win the ‘AI wars,’” McDaniel said this week. “Put another way, companies like … ChatGPT and Anthropic are producing technology that can be, not readily, but within a matter of months, reproduced by the public.”
As a result, free-to-use AI platforms are steadily becoming more powerful as the technology advances, meaning “the economic differentiator of these larger companies are going to diminish,” he said. At the same time, the “really dominant” businesses will be the ones providing the critical infrastructure underpinning this broader AI usage.
“These large data centers are going to be the utilities for AI,” he said. “You will need to use these large-scale utilities to practice AI, even though the … software may be readily available.”
McDaniel also weighed in on the public’s “rejection” of AI, particularly among young people. He referenced a study showing just 18% of Gen Z have a positive outlook on AI, arguing this generation generally distrusts the technology.
He drew a comparison with the evolution of advanced internet search engines over the last two decades, noting Google initially dominated the space when it was introduced with its sophisticated relevance-based search function.
But over the last 25 years or so, this model has shifted into a “bidding war over your attention,” McDaniels noted, as results were monetized to prioritize businesses and others willing to pay for the privilege of showing up higher in a search.
“We will see this in AI,” he said. “AI will not necessarily provide you the best AI services for you as a consumer. The concern is, is that the AI will provide the best monetization of you as a consumer, that works in the best interest of the companies. And we’ve seen this over and over again.”
Even as AI becomes an increasingly important tool, McDaniels warned the “information asymmetry” between users and those controlling the technology could lead to manipulation of the public, including in ways that aren’t readily apparent.
“So this rejection of AI is the fear that the same kinds of services, or the misuse of those services, against the consumer’s best interest will play out in AI,” he said, adding “yes, it will be important for the next generation of users to know something about AI. But we have to be really careful about the hype.”
