Digital algorithms were meant to improve my experience online, giving me a more tailored feed of content (and ads) that matched my needs.
I was not supposed to notice how precisely they were suggesting music that matched my exact mood, videos that appeared right when I was thinking about them, or shows and movies I had not searched for but somehow already wanted to binge.
At first, it felt like a win-win situation.
The platform got more engagement from me. I got a more personalized experience. The internet felt less chaotic and more intuitive.
Turns out, it was not that simple.
The attention machine
Algorithms are sets of instructions that help platforms decide what to show online. They rank, recommend, and filter content based on signals from our behaviour.
Originally, they were built to solve a real problem: information overload.
Early systems worked in a relatively straightforward way. If you clicked on something, you would see more of it. If you engaged with a topic, the system would surface similar content.
But as platforms scaled, their objective shifted. Relevance stopped being the main goal. Retention became the priority.
Time spent, scroll depth, watch time, return visits. These became the metrics that defined success.
From that point on, algorithms were no longer just organizing information. They were optimizing engagement.
And that changed what they surfaced, not only reflecting preferences but reinforcing or even manipulating them.
An invisible negotiation I agreed to
When a product is free, or even when it is paid but attention still carries value, the user effectively becomes part of the product itself.
In that premise, participation means accepting that behaviour will be tracked, interpreted, and used to shape what comes next.
Some days, the feed feels useful. Other days, it feels repetitive in a way that is hard to ignore.
The same ideas, slightly reshaped.
Narrowing personalization
Personalization is usually framed as a benefit. It reduces noise, saves time, and makes content feel more relevant.
And it does all of that.
But it also narrows what is surfaced over time.
The more the system learns what keeps me engaged, the more it repeats variations of the same patterns.
Familiar topics, emotions or debates, even outrage. Discovery slowly becomes repetition disguised as relevance.
Eventually, the feed stops feeling like a window into the world and starts feeling like a reflection of what the system has learned about me.
Holding platforms accountable
Recently, courts and regulators have begun challenging how these systems are designed, particularly in cases involving minors.
Social media companies are being questioned over features that may be addictive by design: infinite scroll, autoplay, push notifications, and recommendation systems optimized for continuous engagement.
The central question is shifting.
If a system is built to maximize time spent, and that design leads to compulsive usage patterns, especially among younger users, where does responsibility sit?
This marks a broader shift in how algorithms are understood. Not only as technical infrastructure, but as behavioural systems with measurable real-world consequences.
The outcomes of these cases are still uncertain. But the direction is clear: design is no longer treated as neutral by default.
Perception and fragmented realities
It is tempting to say algorithms directly cause polarization, but the reality is more indirect. What they consistently do is shape exposure.
Different people can experience entirely different versions of the same issue, not because the underlying facts differ, but because emphasis, repetition, and framing differ.
I am not from Argentina and I generally don’t care about sports, but after Spain won the 2026 FIFA World Cup, I saw how the algorithm amplified commentary portraying the Argentine national team, and to an extent, the entire country, in an overwhelmingly negative light.
Recommendation systems tend to reward content that provokes strong reactions, whether outrage, humour, or tribal loyalty.
As more people liked, commented on, and shared those posts, they became even more visible, creating the impression that anti-Argentina sentiment was far more universal than it may actually have been.
At the same time, false quotes, AI-generated videos, and misleading clips also circulated widely, adding another layer of distortion to an already emotional conversation.
In politics, this produces parallel emotional realities around the same topics. Not just disagreement, but divergence in what feels urgent, what feels normal, and what feels true.
In media and economic terms, attention becomes a scarce resource. And in that environment, speed and intensity often outperform nuance.
The result is not a single distorted reality, but multiple algorithmically shaped realities existing at the same time.
So what do I do about it?
I cannot fully exit algorithmic systems because they are embedded in how I search, watch, listen, and read content.
They are not just part of social media anymore. They sit inside search engines, streaming platforms, maps, news apps, even the way I discover music or decide what to read next.
Even when I try to be intentional, I am still operating inside environments that are constantly ranking, filtering, and predicting on my behalf.
But I can change how I interact with them, in ways that reduce how much they shape my attention by default.
How to build a healthier relationship with algorithms
- Turn off autoplay wherever possible: Stop letting the next piece of content decide for you. This applies to video, music, and podcasts.
- Limit infinite scroll apps with time boundaries: Use screen time limits or app timers for feeds designed around endless consumption.
- Decide your intent before opening apps: Ask yourself whether you need to search for something specific, or just to scroll? And set a time limit.
- Reset recommendations regularly: Clear watch history or reset suggested content to avoid being locked into narrow behavioural patterns.
- Follow sources outside algorithmic feeds: Use newsletters, subscriptions, or direct websites instead of relying only on recommended content.
- Diversify what you consume intentionally: Actively choose content outside your usual interests: different topics, languages, or perspectives.
In the end, the goal is to stop disappearing inside the algorithm and to notice when choice is still present, even in systems designed to make it feel automatic, resisting the parts of them that quietly replace intention with habit.
The more visible that becomes, the more room there is to decide what actually deserves attention.
Oriol Salvador 