Williams on Social Media and Democracy

Dan Williams’s new essay, “Is Social Media Destroying Democracy—Or Giving It To Us Good And Hard?” is characteristically sharp and well-argued. I was going to post a short note giving my initial reactions, but that note quickly grew in length and detail, to the point that I decided it was worth a full post. I reserve the right to follow up with an even longer, more detailed post later on.

One minor point is that I think Dan is overstating my position a bit when he says in the ‘Further Reading’ section of his new essay that I think he’s ‘very wrong’ about social media. As with his previous work, I found lots to agree with in this recent piece. That said, no one will be surprised to hear that I still believe he’s understating the harmful role of social media per se, as opposed to its broader democratizing effects.

A brief thought experiment will help to explain why. Imagine a world in which ordinary people had commensurate influence over public discourse, but without social media. Perhaps there are more frequent, larger, better publicized town halls open to all. Maybe there are also cheap, widely accessible T.V. shows hosted by ordinary citizens, ones that compete effectively with shows hosted by media elites. Would this world experience a profusion of sensationalist / extreme / violent rhetoric similar to what we’ve seen post-the rise of social media?

If so, that would speak in favor of Dan’s argument that the recent rise of right-wing populism and political extremism more broadly is the consequence of a more democratic public square. But if not, it would speak in favor of mine, that social media is the problem, not a broadly right-wing, populist public with a louder voice than before.

There are several reasons I suspect we would not see a similar profusion of extreme politics in this alternative world:

a) alternative means of public influence besides social media don’t allow speakers to be anonymous, meaning they would have to worry more about the reputational effects of what they say, discouraging extreme rhetoric;

b) absent social media, speakers (or ‘content creators’) would not be paid in direct correspondence to the amount of attention they garner;

c) nor would that attention be tracked in such extensive detail, making it more difficult to optimize for;

d) and this world would also lack massive engagement-hungry corporations systematically amplifying the most attention-grabbing views (which tend to also be the most sensationalized and extreme ones).

And that’s just a start. I concede I’ve only made a few suggestive comments about the nature of this alternative world, so it’s hard to say exactly what it would be like. Maybe if ordinary people routinely hosted their own television shows, they would still disseminate extreme rhetoric for the purpose of maximizing engagement. I’m skeptical they would, though, without strong incentives of the kind social media creates.

Even if those incentives cropped up naturally in this alternative world, I don’t see that as a problem for my argument. Strictly speaking, my complaint has always been with the specific design features of actually-existing-social media—design features that might have similar harms applied to different contexts—not with all possible variants of social media, some of which I’m sure would be far more benign, or even beneficial. The way this distinction plays into my argument with Dan is that it suggests it would be helpful for both of us to sharpen what we mean by “social media.” I think it’s probably the case that we both support the existence of something social media-like, but would prefer it be governed differently, likely in a way that involves a greater degree of gatekeeping (I have a post in the works about what this gatekeeping should look like).

I think our real disagreement comes down to the proportion of responsibility for our current political malaise owed, on the one hand, to the objectionable design features of actually-existing-social media (principally those that incentivize users to optimize for attention above all else), and the bare fact of it increasing non-elites’ influence over public discourse. Very roughly, I’d put the ratio of responsibility here at something like 70/30 (favoring design), whereas I imagine Dan would put it at something like 30/70 (favoring non-elite influence).

But this binary leaves something crucially important out. As I argue in “The Case Against Social Media is Stronger than You Think,” it is principally elites, not non-elites, that are responsible for the extreme political rhetoric that has exploded on social media. Yes, smaller accounts make extreme posts too, and sometimes those posts go viral. And yes, they make up the followings of more influential actors like Tucker Carlson and Candace Owens, and transmit those bad actors’ views to others. But it is the Tucker Carlsons and Candace Owenses themselves that are mainly responsible for social media’s inflammatory impact. It is the elite political influencers who are to blame, not (the vast majority of) average social media users.

This is a problem for Dan’s argument, as well as the broader argument that social media has had a democratizing effect on the public sphere. While it surely has at some level, its more consequential effects flow from the elite political influencers it has empowered, only a very small portion of which weren’t elites already, before they gained access to social media.

This suggests a very different conclusion than Dan’s: social media, far from devolving public influence back to the masses, has instead shifted public influence between different groups of elites. In particular, it has favored (in relative terms) those who have for various reasons been denied access to legacy media, perhaps because they have fewer qualms than the average journalist or politician about stoking fear, hatred, and alarm, not to mention outright lying, but also very possibly because of excessive liberal bias in legacy institutions.

That is not to deny that social media has had a democratizing effect. But it is not the same one defenders of the technology have emphasized. It is a stretch to claim social media has given typical members of the public substantially more voice. What it has done, however, is provided a platform for them to more readily amplify others’ voices, namely elite political influencers’, most of them on the populist right. So yes, in that narrower sense, social media has had a democratizing effect: it has unleashed latent public demand for media produced by a different subclass of elites than dominated the pre-social media landscape. But that is very different from democratizing the public sphere as such.

There are two things I want to highlight about this alternative story of social media-induced “democratization.” First, it contradicts Dan’s claim that, until the rise of the internet and social media “information production remained primarily in the hands of wealthy, well-connected elites.” Except at the margins, this is still true. Elites still largely control information production, even if different kinds of elites than before are competing for narrative dominance.

Second, this alternative story has very different implications for the democratic credentials of prospective defenses of digital gatekeeping. It is one thing to prevent typical members of the public from participating in major public fora. It is another thing—arguably a more permissible thing—to limit certain elites’ ability to do so, especially if we have ways of ensuring those restrictions only apply to them; one example of what that might look like is for larger accounts to face stricter content restrictions than smaller ones.

Still, I’m aware that this is an extremely controversial, easily misapplied idea. It is an open question to what extent it is consistent with the values of a healthy democracy to restrict the reach of certain extreme, even anti-democratic views proffered by the rich and powerful. But it is, at least, the right question to ask. Compare it to an analogous question about the public writ large, in particular one that assumes social media has amplified all our voices equally: should we block the masses’ access to social media, given the dangerous new world their voices have wrought? I hope we can all agree that this question is the wrong one, on at least two different fronts.

Edit: I received a helpful comment from Lionel Page (who is an economist at the University of Queensland) that gave me occasion to clarify / expand on some of the arguments above. Thought the exchange was worth reposting in the body of the article (and will update with any future comments from Page / myself).

Lionel Page: Hi Nathan, thanks for pointing me to your post. I don’t think that your two perspectives are really different. Whether the counter-discourse is made by elites or not, it is targeted at the popular preferences that were previously under-represented. I had written the following paragraph for a future post, prior to Dan’s post: “Social media alter the democratic game. They create opportunities for political entrepreneurs to vie for new coalitions and challenge existing ones. They also give citizens a way to seek and supply public justifications for different positions without the intermediation of knowledge institutions that might overly represent elite views. As a result, social media make the marketplace of ideas more contestable.” (Here I refer to the economic notion of contestability for markets.)

Nathan Witkin: Thanks for the thoughtful comment! I think the last passage is mostly spot-on, but would say it omits some considerations that help to bring out where I see Dan and I differing.

One I would underline is the question of to what extent popular preferences are endogenous to the marketplace of ideas. I see online ‘political entrepreneurs’ as powerful shapers of public opinion, so I think it’s misleading to frame our new digital political culture as the product of unalloyed, bottom-up democracy. I still agree with the overall thrust of the idea that social media allows people to seek out perspectives considered beyond the pale in mainstream media, but I think we’re missing a key part of the story if we treat elite political influencers as simply filling preexisting demand, rather than at some level helping to create it.

Some evidence / elaboration on this claim from a previous essay:

“There is strong causal evidence that political influencers have the power not just to reflect, but to shape the attitudes of their followers. In a recent preprint, Rathje et. al. (2025) found in two different field experiments (n = 494, n = 1,133) that unfollowing “hyperpartisan social media influencers” on Twitter (they give the examples of Palmer Report on the left and Breitbart News on the right) “improved [participants’] recent feelings toward the out-party by 23.5% compared to the control group, with effects persisting for at least six months,” (p. 2). This should put to rest the idea that inflammatory political content is a purely ‘demand-side‘ phenomenon. Yes, this demand has always existed, but it is possible for the attitudinal effects of existing supply to increase it even further.”

At a more philosophical level, I am hesitant to accept that social media is “democratizing” the public sphere at all. Given Dan embraces this rhetoric in his essay, I’d say this is also a point where we differ. It’s very much still an unmined vein for theoretical exploration whether enlarging the marketplace of ideas is intrinsically ‘more democratic.’ How do we make sense of this claim if the ideas proffered by new entrants are grossly undemocratic in substance? What does it mean for democracy writ large if the marketplace of ideas becomes more procedurally democratic while becoming less substantively democratic? Is this even a well-formed distinction? We’re still a long way from having hashed these questions out.

Lionel Page: Thanks, Nathan. I think the two dynamics (political entrepreneurs creating content and masses demanding content) are two symbiotic things. Entrepreneurs try to create content that has the maximum chance of success. They innovate, but innovations will work only to the extent that they meet latent preferences in the population. For that reason, entrepreneurs will most often offer the population things they want (e.g. Jobs creating iPods and iPhones). Could they sometimes lead the population astray with bad innovations? Sure, appealing but bad products can be offered, and the same can happen in the marketplace of ideas. It does happen (not just on the counter-elite/populist side, though).