
Commentary

Kevin Frazier, Assistant Professor of Law at St. Thomas University Benjamin L. Crump College of Law since 2023, previously served as a Judicial Clerk to Chief Justice Mike McGrath of the Montana Supreme Court. A graduate of the Harvard Kennedy School and UC Berkeley School of Law, his research delves into the intersection of emerging technology, democratic design, and the law, examining how governance systems adapt to rapid technological advancements. Professor Frazier's extensive scholarship has appeared in numerous journals such as the Rutgers University Law Review, Virginia Journal of Social Policy & the Law, and Notre Dame Journal of Emerging Technology. His ongoing status as a Senior Research Affiliate with the Legal Priorities Project allows Professor Frazier to remain on the vanguard of updating laws and the legal profession in response to emerging technologies. Track his research here or by following him on Twitter: @KevinTFrazier.
To train the next generation of lawyers in the law and practice of voting rights, ballot access, campaign finance, election administration, and democracy protection.
By Kevin T. Frazier
January 25, 1:30 PM
A flood of AI-generated content is spilling into social media, infiltrating newsrooms, and, more generally, polluting our information ecosystem. Akin to the Model-T making it easier for far more people to emit CO2, the rapid spread, adoption, and use of generative AI has enabled millions to disseminate information “pollutants” including misleading content or “misinformation,” false content or “disinformation,” and intentionally harmful content or “malinformation.” It’s true that not all AI-generated content has these characteristics. But even factual AI-content may pose negative consequences—for one, it may displace the journalists in our communities who we’ve come to rely on and trust.
The impending glut of AI-content would not pose as much of a problem if people could easily distinguish between “artificial” and “organic” content. They can’t. It also would not be a problem if platforms and publishers adopted a hard and fast rule against artificial content. They won’t, have yet to do so, or, more practically, can’t, without relying on users to self-identify the artificial content. Finally, this surge would not justify as much concern if AI-generated content was just a fraction of online content. It’s not—it’s becoming much, much more than that. In fact, by 2026 more than 90 percent of online content may be AI-generated.
In short, there’s been a spike in the supply of artificial content in the Marketplace of Ideas. At the same time, there’s been a decrease in demand for “hard” news—news covering “timely, important and consequential” topics. This isn’t hyperbole (trust me, I wish it was). According to the folks at Pew, “[i]n 2016, 51% of U.S. adults said they followed the news all or most of the time. But that share fell to 38% in 2022[.]” These two shifts in supply and demand have moved us further away from a socially-optimal equilibrium in which the public at once demands reliable and accurate information pertaining to civil affairs and a steady supply of such information exists at a low cost.
How best to correct this market failure is a complex task. Hard news is expensive to supply. If demand for such news continues to wane, then news outlets will have no choice but to cut back on investigative reporting or, as discussed below, fold all together. What’s clear is that the status quo—in which social media platforms serve as the primary Marketplaces of Ideas—is not going to result in the necessary market correction.
Social media platforms, keen to keep users on their sites for as long as possible, happily oblige the public’s news avoidance by populating feeds with silly and sensational information. Users don’t protest. In fact, researchers determined that users do their part to fill feeds with anything other than information relevant to democratic discourse; of the news stories shared by users, researchers determined that just 13 percent were “hard news (such as national news, politics, or world affairs).” And, according to Reuters, “[a] growing number of people are selectively avoiding important news” such as reports on geopolitics, pandemics, and housing affordability. Why? Simply put, the news bums people out. Unfortunately, this same news is what we’re supposed to rely on to hold our officials to account, monitor policy implementation, evaluate candidates, and identify problems.
As the supply of distracting information surges and our willingness to pay attention to our democracy wanes, the Marketplace of Ideas is in danger of an irreversible failure. In other words, the spread of AI content has all the makings of a crisis of democracy. This new contagion of mis-, dis-, and mal-information is even more alarming given that we were already in the midst of a similar crisis before ChatGPT, Gemini, and the like spread faster than a new Taylor Swift album. Getting our way out of this tailspin is going to require a massive intervention into the market—traditionally a job left to the government.
The government, though, has few legal means to directly intervene in the market to try to increase the production of and demand for democratically-salient information. Modern First Amendment doctrine significantly limits any effort by the government to favor and incentivize the consumption of certain speech.1Miami Herald Pub. Co. v. Tornillo, 418 U.S. 241 (1974); See Tim Wu, Is the First Amendment Obsolete?, 117 MICH. L. REV. 547, 549-50 (2018), https://repository.law.umich.edu/cgi/viewcontent.cgi?article=1960&context=mlr (“The protection of a healthy speech environment in our times demands a rethinking of what it means to protect the channels of political speech in the Internet age.”); Kevin Frazier, The Right to Reality, LAWFARE (Dec. 21, 2023), https://www.lawfaremedia.org/article/the-right-to-reality (“Though the U.S. Supreme Court has labeled cyberspace as one of the “most important places … for the exchange of views[,]” it has yet to formally declare social media platforms as public forums as defined by First Amendment jurisprudence.”). However, the government is not without any options. As pointed out by Richard Stengel, the government faces no First Amendment constraints when it comes to restricting libel and speech that leads to imminent lawless action. Still, the health of our information ecosystem requires more than attempting to remove toxic content—we must dilute the effects of those pollutants by increasing the supply of reliable, verifiable, and actionable news.
Federal and state governments should explore means to increase the financial viability of local newspapers. Any such assistance could cause a noticeable spike in the supply of democratically-salient news, at least at the local level. According to the Local News Initiative, “[s]ince 2005, the country has lost almost 2,900 newspapers, including more than 130 confirmed closings or mergers over the past year.” Some of these papers cannot be revived, but that doesn’t mean the government should not lend a hand to entrepreneurial folks around the country who are exploring new ways to cover and share local news. By way of example, check out Oregon360 Media, which is based on a crowdsourcing model.
State governments should work on the demand side of the market by making civics a core part of K-12 education. One strategy—employed by two states—is to make high school graduation contingent, in part, on students passing a civics exam. As much as I’m not a fan of teaching to the test, the current state of the Marketplace of Ideas and our democracy suggests that we would benefit from more young Americans becoming accustomed to spending more of their daily attention budget on news of relevance to our democracy.
Finally, and more ambitiously, the federal government should evaluate the creation of “News Exchanges.” Funded with government subsidies, these exchanges would be social media pages, feeds, etc. that are reserved for the distribution of content from verified news outlets. Instead of forcing users to guess whether content is legitimate and sort through mountains of memes, these exchanges would operate as expressways for explicitly democratically-valuable information. Notably, users would be under no obligation to spend time on these exchanges. Instead, the exchanges would merely provide users with the option of directing more of their attention to hard news, an option that may not be available absent News Exchanges or a similar intervention.
A republican form of government cannot work if the people do not stay informed on civic affairs. This has been recognized by individuals such as Thomas Jefferson—who famously declared a preference for “newspapers without government” over “government without newspapers”—and organizations such as the United Nations, which sets forth a right “to seek, receive and impart information and ideas through any media and regardless of frontiers” in the Universal Declaration of Human Rights. To guarantee access to such information in the Age of AI, we must recognize and protect a broader right to reliable and verifiable news or, as I have taken to calling it, a “Right to Reality.”
A looming collapse of the Marketplace of Ideas indicates that now is the time for drastic measures to increase the supply of and demand for hard news and other democratically-salient content. The proposals offered here are not intended to end the debate over what to do but rather to spur critical conversations about the need for an immediate and substantial intervention.
References:
Related