Pipelines to different planets

August 01, 2025

Javiera Scarratt

While we scroll on our phones, a machine learning algorithm picks up our every swipe, tap, and hesitation so it can feed us exactly the content that will keep us on the platform as long as possible. The longer we scroll, the more of our attention can be auctioned off to advertisers, the more the company profits. ‘Filter bubble’ is a term coined by Eli Pariser in 2010 to describe ‘that personal ecosystem of information that's been catered by these algorithms to who they think you are’.[1] Research into this idea has largely focused on the news sources (narrowly defined) that people are exposed to, and has found little evidence that many people are truly trapped in online information bubbles.[2] I suspect, however, that this narrow focus has obfuscated the bigger story here. Most people don’t spend very much time reading the news.[3] But people, and especially young people, do spend a lot of time scrolling through feeds of ‘content’ chosen for them by algorithms, particularly with the meteoric rise of continuous short video feeds like TikTok and its copycats (YouTube Shorts, Instagram Reels, etc.). This content tends to be organised and delivered along well-worn channels, sometimes called pipelines. Through constant population-level experimentation, the algorithms have carved out channels that are extremely effective at keeping people glued to their screens.  

Say the TikTok algorithm learns that you enjoy videos about pottery. It shows you more videos about pottery, yes, but it also knows that people who like pottery are likely to enjoy videos about stained glass making, and people who like both will probably be interested in sewing content. These genres are clearly related, but once the platform knows this much, it can also make other guesses about you based on existing audience trends. You’re probably a woman, probably young, probably left-leaning, probably a feminist. Whether you are these things or not, the algorithm will show you content made from these perspectives. Crucially, it will not primarily be ‘news’ content. It will be videos about sewing that mention the environmental impact of fast fashion or videos of women throwing vases on pottery wheels while talking about their ex-boyfriends’ weaponised incompetence. And when the algorithm clocks that you’re interested in these topics too, you’ll see more of them—now without the bridge of pottery or sewing. Unless you have firm prior opinions about these things, you’re likely to begin believing that the perspectives of these reasonable-seeming content creators you like and identify with are correct, widespread, and mainstream.

The most infamous form of this phenomenon is the ‘alt-right pipeline’, in which young men are algorithmically lured down a rabbit hole of increasingly violent, misogynistic, and racist content. Until YouTube manually intervened in its algorithm to prevent flat earth and QAnon videos from gaining traction, the conspiracy theory pipeline was also known for its radicalising effect.[4] But the algorithmic function underlying these channels is no different from that which drives the above, relatively benign, pottery example, and which drives the online experience of everyone who uses these platforms.

The communities formed by these divergent channels inhabit entirely different online worlds that are almost invisible to each other. I would suggest it is this kind of content delivery system that creates real and consequential ‘filter bubbles’, rather than the possible algorithmic prioritisation of certain news sources. A more apt metaphor might be a clustered content network, in which the pottery, sewing, and art history niches form a cluster with many points of connection between them, but with very few connections to the pro-wrestling or conspiracy niches.  

One important aspect of these clusters is that they’re highly gendered. Showing interest in a male-coded subject will prompt the algorithm to fill your feed with content made for a male audience, and vice versa. Some niches live at points of gender overlap—popular science content, for example, seems to have a fairly gender-balanced audience—but more often a particular community will skew heavily towards one gender. Algorithmic platforms make it probable that, for young people especially, most of our time online is spent immersed in gender-segregated communities, being served content calibrated to appeal to the interests and anxieties of our sex (whether healthy and well-founded or otherwise). Even two people seeing exactly the same content may find it framed very differently. The comments we see, like our feeds, are algorithmically personalised to maximise engagement. In one viral example, a woman saw an Instagram Reel about a man who had gone out golfing and been hours late to an appointment with his girlfriend.^[5] The comments shown to the woman were sympathetic to the girlfriend, pointing out that it was rude of the man not to show up or to let her know he’d be late, along with more extreme comments like ‘he’s replaceable’ and red flag emojis. The woman sent this Reel to her boyfriend, who was sitting beside her. But when he opened it and looked at the comments, he found the opposite perspective: ‘Or you could get your own hobby instead of waiting around for him,’ and ‘God forbid he has a good time’ are examples of the dominant attitude shown to him.

This online segregation of young people has real world implications. One measure seen worldwide is the rise in gendered political polarisation. In 2022, 34% of Gen Z men voted for the Coalition compared to only 20% of women, and data from the 2024 American presidential election showed that while the gender gap of voters aged 35-60 was about 10%, it was well over 20% among 18-25 year olds.[6] But this is only one facet of a much larger phenomenon. Our political preferences don’t emerge in a vacuum; they’re the visible, measurable manifestations of the stories we tell about the world and who we are in it. The content we steep ourselves in profoundly influences the moral and conceptual frameworks we use to interpret the world around us. If these are being shaped more and more by time spent online in single-gender silos, and the flawed ideas we imbibe aren’t being challenged by time spent interacting with each other in the real world, the fault line being formed will have social ramifications far beyond politics. We’re already seeing a decrease in young people dating and getting married, and while the reasons for that aren’t fully known, it can’t be helped by the increasingly incompatible worldviews and value systems being entrenched in us by online communities.

Cross-gender friendships are an important part of the solution to this growing rift. Research suggests that gender-mixed social circles correlate with a lower prevalence of misogynistic attitudes and a reduction in stereotypical views of the other gender.^[7] Churches can play an important role in forging connections across social boundaries of all kinds, including gender, but this requires deliberately creating a culture in which cross-gender interaction and friendship is facilitated and encouraged. Such a culture is far from inevitable—research from the US shows that regions with higher rates of religiosity also show higher rates of social gender segregation.[8] The vast majority of the single-sex schools in New South Wales are Christian schools. Churches and youth groups frequently break into single-gender small groups for bible studies and social activities. Many churches and Christian gatherings self-segregate into gendered groups during social times like morning tea. But the biblical vision of the church is one of unity—neither Jew nor Greek, male nor female, slave nor free—not one of fragmentation. A culture of gender segregation, especially when it is being compounded by gender disconnection outside the church, is unhelpful. Moreover, the church can serve young people (for its own sake as well as theirs) by facilitating the cross-gender relationships and conversations needed to challenge the skewed worldviews and reductive stereotypes engendered by social media. This isn’t an onerous additional activity to add to the church schedule. It just means being intentional about fostering a healthy community of believers, where Christian brothers and sisters spend time together learning to love and serve and respect each another, in recognition of one another’s full humanity, complexity, and dignity as fellow creatures made in the image of God.

Leave a comment

Comments will be approved before showing up.

Name *

Email *

Comment *

---

Also in CASE Online Library