Critique of a Notion of "Filter Bubbles"

What is the source of the internet's apparently great polarization? Uwa, such fake news... such ideology... Oh my, the incels! The Neo-nazis! What could be responsible for this dastardly brainwashing? Is it, perhaps... the queers? I shiver at the notion, but... Yes, surely, it must be... The queers are to blame, I know it...

</end bit>

Anyway, I've been thinking about this popular theory of "filter bubbles" and "echo chambers". There seems to be widespread serious concern that social media algorithms are dividing people through this process. These words were initially defined as technical terms, I think, with very particular meanings. What I'm concerned with here is not their academic use, but their popular use, where the two words are seemingly synonymous.¹

Here is the popular conception, as I understand it:

Social media content delivery algorithms seek to show users things that they like, information from 'their side' which they agree with. The platforms 'echo' back information which fits users' existing beliefs, extremizing them. Alternative takes are 'filtered' out, secluding users in 'bubbles' of their peers, which mutually reenforce one anothers' beliefs, without exposure to outsiders. It is speculated that in the future, as this process is optimized, users will be entirely secluded from the outside world, living only in their own self-reenforcing fantasies.

This is a theory, and it makes a prediction. If everyone is secluded in filter bubbles in this way, what behavior do we expect to see when we go online? Since social media users are allegedly only being served information which matches their existing beliefs, and only interacting with users who agree with them, the theory predicts that interactions you see online should be near-exclusively between like-minded parties, with no significant arguments or disagreements. This is the single prediction that the theory makes: The internet should have no arguments, because opposing views are filtered out.

To anyone who has spent any time on social media, this is a laughably bad prediction. The theory is obviously wrong, and I really don't know how anyone can take it seriously. To believe in this notion of filter bubbles, it seems you would have to be more delusional than the alleged filter bubble victims you theorize.

The internet doesn't hide opposing views from you, creating filter bubbles. No, it entirely destroys the conditions of the existence of such a process. Information disseminates more effectively than ever before, and thus, hiding opposing views is a less practical manipulation strategy than ever before. These kinds of filter bubbles can only exist when there is a serious restriction of information, for example, when everyone is relying on the same news channel to learn about global events. In that case, all information is squeezed through the same system, passed through the same set of hands, which may filter as they please, and thus gain large-scale control over beliefs. With the internet, this is practically impossible. Rather than centralized information systems with centralized authority, there is multiplicity, and thus no one is subject to any particular filtering.

So, what is actually going on? I'll refer specifically to Twitter from here on, since it seems like the best example, but what I'm saying should apply to other platforms just as well. Twitter's business model is to keep you "engaged", to keep you on the platform so they can collect data and show you ads. Twitter does not effectively create engagement by showing users only information they already agree with. It's the opposite: Engagement is produced most effectively by getting people into arguments over every conceivable disagreement.

Here is what is being confused for filter bubbles: The kinds of disagreements which are best for engagement are obviously not careful debate, but quippy one-liners. The kind of alternate perspectives Twitter shows you are not genuine, steel-man versions of their positions, but caricatures. This is how polarization and echo effects are produced: Not by filtering out opposition, but by parodying it, making a fool of everyone to be dunked on for clout, producing selective distrust.²

So. If the apparently common conception of filter bubbles is so obviously wrong, why is it so popular?³ Well, what does the theory produce? Surely, it makes you, the Twitter user, who "engages" in "debate" with alternative views, feel uniquely accomplished. After all, you aren't like those mindless normies trapped in delusional frameworks... No, you're above that, because you interact with people who disagree with you! (By calling them stupid on the internet.) The incentive is to continue "engaging".

For some reason, I have been using Twitter a lot lately. If you haven't noticed, there is a lot of transphobia there. Interacting with these transphobic messages does not create healthy, productive conversations with alternative perspectives. For the most part, these messages are just hate speech. At best, they don't affect me, and at worst they make me (and presumably other trans people) feel depressed, afraid, alone, or, to be blunt, suicidal. Filtering them out has a purely positive effect on the user's experience. Who benefits from these messages being seen? Well, they're good for engagement, aren't they?

If you seriously want to understand contrary positions... read a book. Or, idk, read a blog post, or look up lectures, or a debate between experts, or find someone to talk to seriously... anything but social media. Twitter is obviously not the place to look for genuine theoretical understanding- Its systems function by producing cartoonified belief systems fine-tuned to generate moral outrage and quippy dismissals, reproducing further cartoonified belief systems. Everyone knows this. Sometimes, when one speaks of "filter bubbles" or "echo chambers", this is what one is referring to. But it is important that you do not misunderstand- "engaging" with alternative views on social media is not a defiance of systemic forces, it is exactly the process by which these forces reproduce themselves. And, it's bad for you!

1. As far as I'm aware, the academic use does address the problems I am raising. There is a risk here that I am critiquing a strawman, attacking my own misunderstanding rather than a position anyone actually holds. But, idk, it does seem to me like people think this, and I found this whole line of thought interesting, so I'm writing about it regardless.
2. Again, I am pretty sure this is well understood and accounted for by academic theories of filter bubbles. I thought up this whole post before I started writing, but now I'm thinking I might be in some sense falling victim to the very process I'm critiquing? I am fighting a caricature of the theory of filter bubbles, aren't I? Well, at the very least, I am writing a blog post rather than arguing on Twitter, so surely I haven't fallen for the meme too badly...
3. Assuming I'm not hugely overestimating its popularity.