The algorithms that choose individual-specific content are crafted to maximize the time people spend on a platform. As the developers concede, Facebook’s algorithms are addictive by design and exploit negative emotional triggers. Platform addiction drives earnings, and hate speech, lies and conspiracy theories reliably boost addiction.
Careful studies have shown that Facebook’s algorithms have increased political polarization significantly. Researchers have identified a small group of right-wing personalities — Dan Bongino prominent among them — whose influence on social media played an outsize role in promoting false beliefs about the 2020 presidential election. And witness testimony leaves little doubt that posts on a variety of social media platforms helped provoke the Jan. 6 assault on the nation’s Capitol.
Some people object to reining in social media on libertarian grounds. John Samples, vice president of the Cato Institute, a conservative think tank, says, for example, that government has no business second-guessing people’s judgments about what to post or read on social media. That position would be easier to defend in a world where individual choices had no adverse impact on others. But negative spillover effects are in fact quite common.
When an accident blocks the southbound lanes of a freeway, for example, it also causes long delays in the northbound lanes, because many northbound drivers judge the scene worth the 10-second delay to slow down for a closer look. Yet the cumulative impact of those decisions may be several hours of additional delay for drivers behind them. If drivers could decide collectively, most would surely reject that trade-off. But drivers make such decisions individually, not collectively.
For parallel reasons, individual and collective incentives about what to post or read on social media often diverge sharply. There is simply no presumption that what spreads on these platforms best serves even the individual’s own narrow interests, much less those of society as a whole.
In short, the antitrust remedies under consideration in Congress and the courts won’t stem the abuses that flow from the targeted-ad business model. But a simpler step may hold greater promise: Platforms could be required to abandon that model in favor of one relying on subscriptions, whereby members gain access to content in return for a modest recurring fee.
For those willing to pay the fee, this model satisfies the economist’s efficiency criterion, since they can enjoy unlimited quantities of a platform’s offerings at a zero marginal charge. Major newspapers have done well under this model, which is also making inroads in book publishing. The subscription model greatly weakens the incentive to offer algorithmically driven addictive content provided by individuals, editorial boards or other sources.