AI regulation has been, and continues to be, one of the most pressing policy issues of our time, with tens of millions of dollars and thousands of man hours invested in attempts to tip the scales. As these emerging technologies keep advancing, fears around them grow, and outright bans and restrictions become more enticing to some. This has proven to be a great demonstration of many public choice theory insights, including rent seeking and the alliance of bootleggers and Baptists. But our dialogue around regulation is missing a component of the tradeoff of regulation: institutions.
To quote Nobel Laureate Douglass North, institutions “are the humanly devised constraints that structure political, economic, and social interaction,” constituting “both informal constraints … and formal rules.” These rules are necessary for the functioning of markets and society overall. In this piece, I’ll pay special attention to two categories of institutions: regulatory and cultural.
Regulatory institutions are a subset of the larger body of legal institutions, distinguished mainly by the particular tendencies of regulations. The most important regulatory institution when it comes to emerging technology is the idea of permissionless innovation, which has been developed and promulgated by tech policy scholar Adam Thierer. Permissionless innovation is a regulatory approach to new sciences and technology, and seeks to respond to actual challenges, rather than to head off theoretical ones.
Cultural institutions are complex. The most important ones for this argument are the idea of the bourgeois bargain, dynamism versus stasism, and openness. The Bourgeois Bargain is Deidre McCloskey’s explanation for the world-changing Great Enrichment, and is a characteristic of societies that may not celebrate the power of capitalism, but do leave the capitalists alone so they can do what they do best: make us rich. Dynamism refers to societies that accept innovation and change while still acknowledging there may be costs, contrasted with stasists, who seek to freeze the world as it currently is. And openness refers to societies that enable free discussion, thought, and exchange, allowing for “ideas to have sex” as Matt Ridley so perfectly described.
How does regulating AI alter our institutions and why does this matter? It does so primarily, but not exclusively, by interfering with many of the existing frameworks I’ve already talked about.
When it comes to regulatory institutions, regulating AI preemptively and in response to proposed rather than actual harms threatens the very nature of permissionless innovation. This will become doubly true if regulations are designed in a way that forces developers to get explicit permission before making any or even just certain choices. Given the role Bill Clinton’s embrace of permissionless innovation played in the information technology revolution, an abandonment of this framework could rob us of the countless benefits AI is poised to bring.
Cultural institutions are built, in large part, around cultural consensuses and trends that don’t exactly change overnight, but can be redirected. Regulating AI violates our existing bourgeois bargain, interfering with what the capitalists do (and risks they must take) while still expecting them to produce the same results. This endangers the very economic activities that have, and will continue to, produce much of our growth. Regulating AI supports and even strengthens stasist mindsets by showing government support for stasist attitudes and fears. People with dynamist aspirations, from scientists to inventors, will be discouraged, and may reduce or outright stop their aspirations out of an unwillingness to waste time on things that will be shut down. And finally, regulations can create a culture of fear, and an unwillingness to discuss ideas that may be deemed pointless or outright unviable.
While I’ve given specific examples of ways regulations will shatter our institutions, it is helpful to restate this in a more broad way. These institutions matter because they are what empower the dreamers and workers who make our world better. This isn’t just the scientists and inventors, it’s the investors, the venture capitalists, the consultants who make things happen. And these changes will affect future generations too. An aspiring scientist may decide it isn’t worth it, a potential tech entrepreneur may switch to something less productive, and kids may choose to pursue less ambitious dreams. To paraphrase Bastiat, there is both a seen and an unseen. And when it comes to the interplay between regulations and institutions, much of what is lost is unseen; we will never know what could have been had the dreamers and doers been left alone.
Which regulations may be necessary is a discussion I’ll leave to much more seasoned and skilled writers and thinkers. Instead, I merely remind us that when considering what we’re doing, we should make efforts to consider all the effects of our choice, rather than just the first or most obvious ones. Once we imagine all the unseen that could be lost, the true consequences of regulation become more apparent.
Share This Article

Post on Facebook

Post on X

Print Article

Email Article