Who Drives Political Information?

In my first post I laid out an example of the kind of argument the “Discourse Machine” analogy allowed me to make, with the promise to come back and support that argument in a later post. There are a few Kantians in my grad program, and Kantians generally think that promise-keeping is a categorical imperative, so I’m going to avoid getting on their bad side by following through and talking about who, exactly, is the primary driver of political information.

You can go look at the way I laid out the argument in that post, but to save space here, it boils down to the claim that the state of our epistemic environment is driven more by the demand side than the supply side. The point is to think about how we get people like Alex Jones, not just as people with crazy opinions, but as a successful and influential media operation. (Well, successful up until the point he filed for bankruptcy after getting sued for defamation, I guess.) Is it because Alex Jones is an especially talented evil genius who hoodwinked millions of basically innocent and normal people into becoming his audience? Or is it because lots of people like and want Alex-Jones–style content, and he was reasonably good at providing it? 

I don’t think the answer is ever completely one or the other, at least not in plausibly realistic situations. Supply and demand always interact and provide feedback. But we can think about certain conditions where the supply-side or demand-side has more weight, sometimes quite a lot more. Let’s propose two hypothetical cases.

MONOPOLY MOGUL: an extremely wealthy media mogul has convinced the government to provide him with the only license to operate a single television news channel, in a world where television is the sole method of receiving live information (i.e. no radio or internet). 

UNIVERSAL LIVESTREAMING: Everyone in this futuristic world is connected in a Black Mirror-style technological setup where they are constantly broadcasting to anyone who cares to view their channel. There are no entry costs, no censorship, no content moderation, no limitations at all. People watch whoever they want to watch, all the time, and people can do or say whatever they think will get more viewers. People are incentivized to get as many viewers as possible (imagine this is the only way this society earns money).

These cases are obviously extreme examples, but let’s work with them for a moment. In MONOPOLY MOGUL, the news supplier has all the power. He can’t necessarily present anything he wants, because it’s possible for people to just turn off the television and not get any news at all if his programming is sufficiently unpopular. But lots of people will want to watch the TV and see the news, so he has a lot of leeway to control the information people get with such a captive audience. In UNIVERSAL LIVESTREAMING, the demand-side has all the power. If a viewer doesn’t like what another person is doing or saying, they can go browse for some other kind of content. It’s very difficult to shape the information environment because if you do anything that goes against what your watchers like and enjoy, you risk them ‘turning you off’ and watching someone else. This is what sometimes gets called ‘audience capture’ and it’s a powerful force in the real world, not just my thought experiment.

In real life we exist, at different historical moments, in a spectrum between the two cases. The peak of legacy TV and newspaper media, dominated by the major channels and papers of record, had some features more in common with MONOPOLY MOGUL. Those outlets could carve out slightly different audience niches (Fox News and the Wall Street Journal could cater more to conservatives), but there were limits to how sharp that slant could (profitably) be. But where that slant existed, was it due to the ideology of the owner, or that of the audience? Both may have played a role, but this study provides what I think is pretty compelling evidence that “an economic incentive for newspapers to tailor their slant to the ideological predispositions of consumers” explains much more of observed slant than the ideology of owners. That study provides some important caveats, though, including the idea that individual owners and reporters can impart significantly more slant in their coverage of issues they are very concerned with, but readers are generally less aware of.

The model for legacy media, I think, is roughly this: broad ideological slant has to align with that of expected consumers (I’m not going to be very successful running a liberal local newspaper in a deep red area); but owners and journalists can exert more control at a lower level, like raising or lowering the salience of specific issues within that broad ideological slant. Over time, a sufficiently dedicated effort could create gradual shifts in the particulars of that ideology’s beliefs, important issues, and favored policies. One could imagine a concerted effort by Fox News in 2016 to, say, discredit Donald Trump and rally support around another Republican candidate for reasons that would appeal to its conservative viewers. One could even imagine such an effort working!

That’s very different from the model for social media, however. That model looks a lot more like UNIVERSAL LIVESTREAMING, with similar risk of audience capture. There are lots of complicated reasons why that’s so, but I’ll highlight three of them here.

1. Extremely low barriers to entry. If I want to write for The New York Times or The Atlantic, I have to go and actually convince somebody there to let me write for them. They have to believe that my writing would add value or interest, and that the things I want to say are the things they want their subscribers to read. By contrast, launching this publication cost me about three hours of setup work (a less computer-illiterate person could probably do it faster), whatever time it takes me to write each post, and zero money. Hosting the Discourse Machine podcast costs, like, a hundred bucks. Of course, that’s no guarantee that any given product will get more than a handful of readers - very few people ‘make it big.’ But since the opportunity cost for trying is so very low, there’s a very high probability that someone will try to fill any imaginable content niche.

2. Revenue models. It’s not like advertising hasn’t been a part of the financial model for media for a very, very long time, but it wasn’t necessarily as all-consuming as it can be now. Clicks, views, and listens drive money, and those can come from anything ranging from a committed base audience to endless cycles of rage bait or spamming AI-generated slop. Accuracy, quality, and truthfulness might be compatible with certain revenue models, but they aren’t exactly inherently successful strategies. 

3. Discomfort with intellectual discomfort. Closely related to the idea of confirmation bias, we humans often don’t like the uncomfortable experience of confronting evidence or arguments that suggest we ought to change our minds on something. Sometimes that’s because of internal reasons, like having to admit that I was wrong - maybe really wrong - about something previously; sometimes that’s because of external reasons, like the fear of losing my connection to or approval from a group I’m currently a part of. Either way it’s not exactly fun. Here’s the neat trick now: whatever you think, or want to think, you can go find somebody that passes for credible, or at least popular (so you’re not one of 17 people watching some crank YouTube channel) that you can use to make yourself feel justified and comfortable. This is even better than just ignoring information that makes you uncomfortable, instead you can tuck yourself into a warm comfortable blanket made just for you out of whatever fabric of incoherent nonsense is most appealing. ‘Copium’ comes in more flavors than vapes do, and is sold by almost as many different vendors.

I’m not quite so cynical that I don’t recognize that this demand-side influence can have some positive outcomes too. Certainly there are really good independent media outlets and individual writers that have a greater degree of reach and influence, precisely because of this model, than they might otherwise have had. On the flip side, there are lots of ignorant bad-faith actors who fail miserably even in this permissive information ecosystem. But I can’t help but notice the extent to which good, serious, thought-provoking people are at a competitive disadvantage compared to the grifters, the wildly ignorant, and the plain crazy. The net effect is a lot more stuff from the latter categories spreading faster and farther than it otherwise would have. The net effect of that, I think, is a substantially worse world. 

I like having this little publication. I enjoy writing, and I enjoy sharing some of the things I work on academically with a broader audience. I appreciate everyone who reads it. But if I could pick between keeping this around and eliminating the incentives and structures that allow people like Alex Jones or Steve Bannon to thrive and spread their message? I’m burning Discourse Machine to the ground and salting the earth after it’s gone in a heartbeat, baby. 

I can’t do that, of course, so I’m going to keep plugging away here, like the guy in the meme.

Here I am, participating in an information ecosystem that I think we should improve somewhat! 

The reason I think this is important to talk about is because a lot of us really like to blame other people for the information space being the way that it is. Lots of conservatives blame the liberal coastal elites that make up the mainstream media. Leftists and those liberals who tend towards materialist explanations blame capital and class. Other liberals blame individual bad actors, overstating the malign cleverness of conspiracy peddlers, the power of provocateurs, and the influence of foreign countries. Those things play a role, but the role they play is on the margins. Right now, to see the real culprits, we ought to look squarely in the mirror.