Publishing Without Romance

Happy New Year!

What follows is a theory I’m not sure I endorse. It may be too cynical to be true.

In 2018, I was fortunate to audit Saul Levmore’s Public Choice at the University of Chicago Law School, a hotbed of law and economics. One insight I took away: When considering any regulation or policy, we can ask, whose interests does it serve? Who stands to benefit from shaping market forces and social behavior more broadly?

Sometimes, the answer is the public interest. Sometimes it’s simply what voters want, and in a democracy, lawmakers ought to honor the people’s will. In most such cases, which one hopes would cover most policymaking, rules coordinate behavior and align incentives with desired social outcomes. The core insight of public choice theory comes from economics: incentives matter. People respond to incentives, so even legislation aimed at promoting the public interest must consider how individuals will actually behave. Most people are not guided by the public interest in their everyday behavior, and they are unlikely to understand how to promote it—hence the need for mechanisms of social coordination.

Many public choice theorists believe that markets generally offer the most efficient mechanisms for promoting social welfare. However, aside from the most hardcore libertarians, most people believe that some regulations are necessary to prevent, if not market failures, at least antisocial and rights-violating behavior, or to deliver certain goods and services. Markets can’t solve crime, and even most economists agree they cannot satisfactorily provide public goods such as national defense, streetlights, or clean air. Public goods are available to all, such that normal market competition fails to price them adequately. Usually, they need to be supported collectively through taxation. Nevertheless, it is worthwhile to ask who stands to benefit from certain rules, for not all rules serve the public interest. This approach can illuminate ostensibly innocuous or even beneficial policies that, upon closer examination, prove to be misguided. It may sound cynical, but cynical behavior requires a cynical analysis.

I aim to apply this framework to current debates about AI tools, specifically large language models and image generators, in academic research. The leading journal Ethics recently implemented a new AI policy clarifying the scope of legitimate AI use in authoring submissions. The journal’s editors should be commended for attempting to address a genuine problem: the misuse, overuse, and abuse of AI tools in academic research. ...