Starving for The Feed: The Rise of AI-Curated Social Media
Stalker-like social media algorithms will do anything to keep us engaged. Why does everything have to be relevant?
The concept of an algorithmically curated social media feed is quite perverse if you think about it. It’s a habit-forming vice that hyperanalyzes your activities and spits out reinforcements. Anything to keep you endlessly scrolling and seeking dopamine hits with likes/reactions and comments. It’s subtly controlling, like an abusive relationship where one party restricts the other’s access to external information.
The slow integration of algorithmic feeds since the 2010s has made us more dependent on a pre-filtered set of news and information, rather than a chronological approach. Facebook pioneered this model in 2011, followed by Twitter and Instagram in 2016. These platforms had already attracted scores of global users but needed to keep them engaged — and thus consuming ads — for several hours a day. Enter the psychology-hacking strategy of projecting the world each unique user wants to see. Instead of a real-time chronology of events as they happened, users received tailored content based on their activities.
Most people never opted out of this new design. After years of fine-tuning, today’s well-curated feeds only need two to eight seconds to capture our attention. This quickly adds up, as the average person spends about 2.5 hours a day on social media — 38 days per year.
The Feed keeps us in a perpetual state of reaction, targeting our instinctual and sometimes trivial human behaviors. For passive users, that might be confirmation bias or knee-jerk reactions based on topics they dwell on the most. Active posters have a whole different response: Heavy social media users grow desensitized to likes and comments over time, and their routine becomes more automatic.
Research is mixed on the real consequences of algorithmic feeds. A series of studies on the 2020 U.S. elections found that they have virtually no impact on polarization. In one study, participants were given a chronological feed on Facebook and Instagram, featuring moderate friends and ideologically mixed sources. Over a three-month period, this change did not significantly influence their levels of polarization, political knowledge, or other attitudes. However, it notably decreased their app activity as they sought alternatives. Wired reported the numbers: Instagram users increased their time on TikTok by 36% and YouTube by 20%. Facebook users moved to Reddit (52% more time) and YouTube (21%).
Chronological feeds are boring now that we know better. It would be interesting to see the same study being conducted over a longer timeframe. Would users eventually come crawling back? Tech giants appear unwilling to find out, understandably, setting the algorithmic feed as the default in most apps: X automatically ranks posts by performance, but users can toggle to a separate “Following” page for chronological posts by accounts they follow. TikTok’s “For You” page is also algorithmic, but you can still view linear posts among users you follow. Facebook and Instagram also sort by relevance, though both provide optional time/date filters. LinkedIn lets users filter recent posts. YouTube’s homepage is full-algorithm, but the subscription page gathers real-time posts from the channels users subscribe to.
Regulators in Europe are now mandating new approaches via sweeping data laws. Last year, the European Union’s Digital Services Act came into effect, forcing major platforms to give users more control over their social media experience. Some have already adjusted their settings accordingly for European users. Meta re-introduced chronological feeds for its Stories, Reels, and search features on Instagram and Facebook, which together total 505 million active monthly users in the EU. TikTok, which has 134 million active monthly users in Europe, now lets users turn off personalization in their “For You” and “LIVE” feeds.
The law doesn’t require platforms to fully disclose their recommendation algorithms, but they are obligated to provide the main metrics of recommending information to users and any options available to change those parameters. Sites classified as “very large online platforms” (Facebook, X, etc.) must always provide at least one alternative option that isn’t based on profiling users for each recommender system. However, this option isn’t required to be the default setting.
Customization options are a positive step in giving users more control, but we should also rethink the broader paradigm that led us here in the first place. We’ve been groomed to expect relevance, rather than determining it for ourselves. The standard has already been set over the last decade. Users would likely find it boring now that they’re used to the contrary. Conversely, by submitting to the algorithm’s whims, our unconscious habits and influences are no longer private to us. Worse, intermittent reinforcement keeps the experience gamified, increasing the likelihood that we’ll be hooked for longer.
We are sacrificing our independence to a psychology-targeting machine, which tracks, logs, and analyzes our behaviors and attention. Our data quickly gets pawned off to advertisers. The Feed is stalking us and extracting value accordingly, prioritizing certain posts and ads based on the data we mindlessly feed it — for free. We are livestock to The Feed.
It is unsurprising that the robotic Silicon Valley Tech Bros had no social foresight in creating their hype/doom machine. They always saw us as precious resources to be mined, while simultaneously claiming the fake mission of “connecting humans.” What a racket.
But let’s be practical here: They have a fiduciary responsibility to capitalize on the business potential of a curated feed. Attention is an invaluable advertising asset. The soaring value of data fuels this $231 billion mega-market. Social media advertising generated $207 billion in revenue last year, with a combined 4.9 billion users.
In fairness, the curated feed effectively does the job it’s designed to do: keep us engaged by presenting the information we’re most likely to interact with. It’s hard to resist that pull. I personally benefit from the situation. Readers are more likely to find my articles through these very same algorithms. Roughly half of U.S. adults get their news from social media. As another silver lining, it’s actually good that narratives spread fast and collapse hard. The masses form a consensus and then proceed to meme the topic to oblivion. There’s always someone to call bullshit on everything. This is a healthy counterbalance.
Still, prioritized content creates the illusion that every news story is relevant to the average person. Staying informed about global events is good as long as you don’t get unhealthily fixated on something that has no tangible effect on your everyday life or work. It’s easy to fall into this trap.
In general, we should be skeptical of anything that manipulates our perception and understanding of information, especially ones with addictive or compulsive side effects. Algorithmic feeds do exactly that, while simultaneously enticing us to waste years of our lives scrolling, watching, posting, liking/reacting, and scrolling even more until we’re old and our arthritis sets in. As I covered in an earlier post about productivity apps, 2.5 hours a day on social media translates to 5–7 years over the course of a lifetime. You can’t get that time back.
Would we be better off by not giving in to The Feed? In taking the extra step of seeking out information directly, we’re forced to focus on what’s important. Instead of getting lost in the AI-organized forage of life updates, food photos, and other posts we’re likely to forget, we could control our experience organically and, perhaps, be more mindful of how much time we’re spending on these platforms.
This post originally appeared in my Substack newsletter. Subscribe for more monthly essays.
Note: This is an opinion essay. Much like a traditional newspaper column, my newsletter is a side channel to voice my personal views and observations. It’s separate from my main gig as a journalist/reporter. You can find my journalistic work here.
Read More:
- (Counter)Productivity Apps — With excessive record-keeping and needless subtasks, productivity apps offer a convenient path for distraction, simultaneously draining our time and executive planning skills. Are they worth it?
- Immortal Data — Growing old in the modern world means accumulating massive volumes of data over decades. Each file saved and stored joins an ever-growing graveyard of digital assets with nowhere to go when we die.
- The Next Generation’s Problem 🔒– In our rush to integrate Gen Z into the digital world, we forgot to teach them legacy tech skills. Is this an oversight or a gift in disguise?
