AI and Social Media: Are We Controlling Our Feeds or Are They Controlling Us?

Are we being controlled by our feed? That is a question worth asking in today’s digital space, where people are spending hours consuming content. I bet a large chunk of Instagram users will nod a hesitant yes to this.

This is not a theory anymore. We are well aware of this fact, and yet we let it control us. Why? Because we are addicted to the fractions of dopamine we get from a 30 second reel that feels relatable, funny, or valuable to us in that exact moment.

Nick Clegg, the president of global affairs at Meta, has said it plainly: Meta uses AI to shape what you see on Facebook and Instagram. People assume they are in control of what they see because they are the ones scrolling, tapping, and choosing to follow certain accounts. But that assumption falls apart the moment you notice how quickly the algorithm catches up to you.

We are shown what we relate to, because that is how the algorithm works. You can see it happening in your own feed. Every time you develop even a passing interest in a new topic, your feed shifts drastically, and Meta adds significantly to making you obsessed with that topic. Is it just coincidence? Not according to Clegg. He explains it this way: our AI systems predict how valuable a piece of content might be to you, so they can show it to you sooner.

So we are shown what we relate to the most, and shown it fast. And before we have the time to think, introspect, or change our stance on something, we are bombarded by one narrative well beyond the point of return. We might be the ones who initiate the interest, but Meta is the one that takes over. It turns a passing hobby into an algorithmic loop that we scroll through for hours, days, and weeks.

Take someone who holds a strong political or social ideology. He engages with a few posts that align with his views, likes them, comments on them, maybe lingers a second longer than usual. The algorithm registers this as a signal of value and responds by feeding him more of the same, then more of the same again. Within weeks, his feed is an echo chamber built entirely around that one worldview. He is no longer seeing the range of opinions that exist on the subject, he is seeing a curated loop of agreement. And because this is all he sees, he begins to believe this is what the majority thinks, this is simply how the world is. He is not wrong to trust his own eyes, the problem is that his eyes have been fed a narrowed version of reality dressed up as the whole picture.

This is the quiet cost of relevance. The algorithm is not trying to inform him or challenge him, it is trying to keep him engaged, and agreement is far more engaging than friction. So it keeps showing him a smaller and smaller slice of the world while making that slice feel like the entire world. He is devoid of the possibility of encountering a different argument, a contradicting fact, or simply a different way of seeing things, not because those things do not exist, but because the algorithm has no incentive to show them to him. His world has shrunk, and he does not even know it, because shrinking it was never announced, it happened one relevant post at a time.

This is where the real question of control comes in. Control implies the ability to pause, reconsider, and choose otherwise. But when a system is built to predict what will hook you before you even know you are hooked, the space for reconsideration shrinks. You are not choosing content anymore, you are confirming a prediction that was already made about you.

And this control does not stay confined to what we see, it seeps into how we think and feel. When a person is shown only content that mirrors his own beliefs, his confidence in those beliefs hardens into certainty, and certainty leaves little room for empathy toward people who think differently. This is how polarisation deepens, not through one big argument, but through thousands of small, personalised feeds pulling people further apart from each other’s realities. There is also a quieter psychological toll, the constant confirmation of what we already believe can breed a false sense of consensus, making disagreement feel not just wrong but almost unnatural. Add to this the compulsive nature of the scroll itself, engineered around the same reward loops as a slot machine, and you get a mind that is not just narrowed in perspective but also increasingly anxious, easily agitated by opposing views, and quietly dependent on the next hit of validation the feed promises to deliver.

None of this means the technology is inherently sinister. Prediction is what makes these platforms useful. It is why your feed does not feel like a random pile of noise. But usefulness and manipulation are not opposites, they often live in the same feature. The same system that saves you the effort of searching for what you like is the system that decides how much of your day gets handed over to it.

So are we controlling our feeds, or are they controlling us? Honestly, it is probably both, just not in equal measure. We hold the first move, the initial click, the first search, the first follow. But everything after that is shaped, accelerated, and amplified by a system built to keep us there. The illusion of control lies in that first move. The reality of control lies in everything the algorithm does with it afterward.

Maybe the more useful question is not who is in control, but how much of that control we are willing to take back. Awareness is the first step, but awareness alone does not switch off a dopamine loop that has been engineered for years to be irresistible. What might actually help is treating our feeds the way we treat any other habit we want to manage: with intention, with limits, and with the honesty to admit when a 30 second reel has quietly turned into a three hour scroll.