Remember: Regulation inevitably stifles innovation.
Unsurprisingly, the Biden administration is trying to capitalize on the popularity of artificial intelligence (AI). First, with its AI Bill of Rights and now with an executive order establishing regulations on AI developers. It even comes with a website! How useful and convenient, right?
Unfortunately, this recent action is neither useful nor convenient. If they really wanted to foster the development of AI in a way that benefits consumers, they would leave it be, allowing AI companies to make their own business decisions.
However, an unregulated market does not sit well with Biden’s constituents. “The state of AI is like the wild west!” they might say. “There are no rules! We need to do something about this!” (Nevermind the fact that the Wild West was not so wild). This is exactly what they said about net neutrality and the internet. Remembering back to 2018 when net neutrality was repealed yields memories of doom and gloom predictions about what the world would look like post-net neutrality. Fortunately, and of no surprise to free marketers, most of these predictions never came to pass. Let’s briefly examine the net neutrality phenomenon and see if there are any insights we can apply to AI regulation.
According to Ajit Pai, the Federal Communications Commission (FCC) chair who presided over the decision to repeal net neutrality, the pre-net neutrality internet was “lightly regulated” and “embraced a market-based approach.” $2 trillion of network investment came out of this unregulated framework. The internet was doing quite well, then, much like the Biden administration’s actions against AI, Obama urged the FCC to regulate the internet. The FCC chairman followed Obama’s demands, and Ajit Pai dissented, warning of the negative ramifications.
As Ajit Pai writes, “That worry was well-founded. Following the FCC’s decision, and for the first time outside of a recession, independent studies showed that investment in broadband infrastructure declined.” Given this dismal reality, we are lucky for the Trump administration having abolished these rules. Ajit Pai summarizes the consequences of repealing net neutrality.
“The evidence is indisputable today that in the five years since the FCC’s decision to repeal net-neutrality regulations went into effect, American consumers are benefiting from broadband networks that are stronger and more extensive than ever. According to independent measurement service Ookla, average fixed broadband speeds in the U.S. are 287 percent faster today than they were in June 2018 (269.28 Mbps download speeds today versus 93.98 Mbps in 2018). Average mobile broadband speeds have increased even more, at 570 percent (156.51 Mbps versus 27.47 Mbps). Millions more Americans have access to the internet today compared with 2018, thanks in large part to private investment in digital infrastructure.”
Ajit Pai continues outlining the positive consequences:
“[C]onsumers are benefiting from more choice than in 2018. Indeed, competition has not just increased but transformed since then. Residential fiber deployment hit an all-time high last year. Wireless companies like T-Mobile are providing high-speed 5G fixed wireless service to millions of customers. Companies like Starlink are launching low-Earth-orbit satellites to support residential-broadband service, especially in harder-to-reach rural areas. And cable companies are expanding their footprint and upgrading their systems to enable much faster speeds.”
The bottom line is that an unregulated internet or, as some might call it, “a wild west internet,” was actually beneficial for consumers rather than detrimental as the Obama administration alleged.
Today, we face a similar choice. Do we want a “wild west” in the market for AI or a market regulated by the government? Instead of falling for rhetoric, we should actually look at history and the executive order itself and tease out the consequences. History is on the side of the free market, but what of the executive order?
The executive order provides for regulations such as anti-discrimination barriers, data privacy guarantees, notices that AIs are being used, and the choice of consumers, such as in the case of customer service, to opt out of using an AI in favor of a human. All of these sound nice, but all will increase the cost of using AI.
This will result in less investment and development in AI. This technology, which has taken years to be developed into a commercializable product, is now being stamped down by the government. To think that this will “harness the benefits of AI” as the Biden administration claims is a joke.
Moreover, these regulations will benefit bigger companies over their smaller competitors. Larger companies are more capable of hiring a human for customer services in addition to an AI. Smaller companies that currently use an AI for customer service will have their costs increased and perhaps they will even be driven out of the market altogether. Some companies may not even see the light of day.
Given these impacts, it’s likely these regulations are at least in part an example of regulatory capture. This occurs when large, incumbent firms are in a position to influence regulations that were originally intended to constrain them. These firms can utilize regulation to harm their competitors, thereby increasing their own profits. This can be a powerful tool for understanding why regulations become law and how regulations can harm, rather than enhance, competition. This recent executive order regulating AI is no exception.
Ultimately, this executive order will increase costs. Much like how net neutrality affected investment in the internet, this executive order will lead to less investment in AI. Additionally, this regulation will harm smaller competitors relative to larger, incumbent firms. The end result is that consumers have access to less of the benefits from AI than they would have otherwise. This will ripple through the market, affecting all firms that use AI, which presently turns out to be a lot of companies.
If a future Republican administration repeals these regulations, we will undoubtedly hear opining from progressive politicians and the corporate media about the dangers of an unregulated AI market, but when that happens, let the example of net neutrality and the lessons of regulatory capture guide you.
Benjamin Seevers is a Mises Institute Fellow and holds a BA in economics from Grove City College. He will begin his PhD in economics at West Virginia University in fall 2023. His research interests include private governance, public policy, and libertarian ethics.
This article was originally published on FEE.org. Read the original article.