{"id":503,"date":"2026-08-10T12:59:30","date_gmt":"2026-08-10T12:59:30","guid":{"rendered":"https:\/\/alexpjose.com\/?p=503"},"modified":"2026-08-10T12:59:31","modified_gmt":"2026-08-10T12:59:31","slug":"blog-1","status":"publish","type":"post","link":"https:\/\/alexpjose.com\/?p=503","title":{"rendered":"BLOG 1"},"content":{"rendered":"\n

What Happens When an Algorithm Decides What You See?<\/h1>\n\n\n\n

Have you ever searched for one video on YouTube and, somehow, ended up watching ten more about exactly the same thing?<\/p>\n\n\n\n

You watch one video. Then another one appears. Then another.<\/p>\n\n\n\n

After a while, you start wondering: Why am I suddenly seeing so much of this?<\/em><\/p>\n\n\n\n

The simple answer is that you didn’t choose all of those videos.<\/p>\n\n\n\n

The algorithm did.<\/p>\n\n\n\n

Every time we use Instagram, YouTube, TikTok or X, we leave small clues behind. What we watch. What we skip. What we like. What we share. How long we stay on something.<\/p>\n\n\n\n

The system learns from these actions and uses them to decide what to show us next.<\/p>\n\n\n\n

It sounds harmless.<\/p>\n\n\n\n

But what happens when the things an algorithm chooses for us are not just funny videos or football highlights, but news, politics and information about the world?<\/p>\n\n\n\n

We don’t all see the same internet<\/h2>\n\n\n\n

Imagine two people opening X at the same time.<\/p>\n\n\n\n

They could live in the same city. They could be the same age. They could even be interested in the same political issue.<\/p>\n\n\n\n

But their feeds may look completely different.<\/p>\n\n\n\n

One person might see posts from journalists.<\/p>\n\n\n\n

Another might see political activists.<\/p>\n\n\n\n

Someone else might see angry arguments about immigration or crime.<\/p>\n\n\n\n

The platform is the same.<\/p>\n\n\n\n

But the version of the platform each person sees can be very different.<\/p>\n\n\n\n

This is why recommendation algorithms have become an important subject of research.<\/p>\n\n\n\n

And a recent experiment on X gives us a good reason to pay attention.<\/p>\n\n\n\n

What happened when researchers changed the feed?<\/h2>\n\n\n\n

In a 2026 study published in Nature<\/em>, researchers conducted an experiment with 4,965 active X users in the United States<\/strong>.<\/p>\n\n\n\n

For seven weeks, some users were given X’s algorithmic \u201cFor You\u201d feed, while others used a chronological feed that showed posts in time order.<\/p>\n\n\n\n

The researchers then compared what happened.<\/p>\n\n\n\n

The result was surprising.<\/p>\n\n\n\n

Among people who had been using the chronological feed, switching them to the algorithmic feed shifted some of their political opinions in a more conservative direction.<\/p>\n\n\n\n

They were also more likely to follow conservative political accounts.<\/p>\n\n\n\n

The researchers found that the algorithm itself was showing more conservative content than the chronological feed. (Nature<\/a>)<\/p>\n\n\n\n

That doesn’t mean the algorithm suddenly turned everyone into a conservative.<\/p>\n\n\n\n

In fact, the study found no significant change in people’s basic party identification or overall affective political polarisation<\/strong>.<\/p>\n\n\n\n

That’s an important detail.<\/p>\n\n\n\n

The researchers found changes in some political attitudes, but not a complete transformation of people’s political identities.<\/p>\n\n\n\n

And that makes the story more interesting.<\/p>\n\n\n\n

So, can an algorithm really change our minds?<\/h2>\n\n\n\n

Maybe.<\/p>\n\n\n\n

But we should be careful.<\/p>\n\n\n\n

Another large study, published in 2025, gives us a reason not to make dramatic claims.<\/p>\n\n\n\n

Researchers conducted four experiments involving almost 9,000 people<\/strong>. They created a YouTube-like platform and manipulated the recommendations people received to create different kinds of \u201cfilter bubbles\u201d and \u201crabbit holes\u201d.<\/p>\n\n\n\n

They found that these changes had limited short-term effects on people’s political opinions<\/strong>. (Harvard Kennedy School<\/a>)<\/p>\n\n\n\n

So, if an algorithm shows you more political content from one side, it doesn’t necessarily mean you will immediately change your political beliefs.<\/p>\n\n\n\n

But perhaps that isn’t the most important question.<\/p>\n\n\n\n

What if the algorithm doesn’t change your mind?<\/h2>\n\n\n\n

Think about your own social-media feed.<\/p>\n\n\n\n

If you keep seeing the same kind of content every day, you may not believe everything you see.<\/p>\n\n\n\n

You may disagree with some of it.<\/p>\n\n\n\n

But you become familiar with it.<\/p>\n\n\n\n

You learn the names.<\/p>\n\n\n\n

You recognise the arguments.<\/p>\n\n\n\n

You start knowing which issue people are angry about.<\/p>\n\n\n\n

At the same time, someone else may be seeing something completely different.<\/p>\n\n\n\n

After months of this, the two of you may sit down and have a political conversation and realise something strange:<\/p>\n\n\n\n

You are not just disagreeing about the answer.<\/p>\n\n\n\n

You have been seeing different things all along.<\/strong><\/p>\n\n\n\n

And this may be where the real power of recommendation algorithms lies.<\/p>\n\n\n\n

An algorithm doesn’t necessarily have to tell you what to believe.<\/p>\n\n\n\n

It may only have to decide what you keep seeing.<\/strong><\/p>\n\n\n\n

India is part of this story too<\/h2>\n\n\n\n

This isn’t only an American concern.<\/p>\n\n\n\n

A 2026 study looked at 825 young Indian voters<\/strong> and examined the relationship between politically similar social networks, algorithmic recommendations and political polarisation.<\/p>\n\n\n\n

The researchers found that algorithmic curation and politically similar networks were linked with stronger echo chambers and more polarised positions. (ScienceDirect<\/a>)<\/p>\n\n\n\n

This was a survey-based study, so it cannot prove that algorithms directly caused the polarisation.<\/p>\n\n\n\n

But it raises an important question for a country where social media has become a major part of political communication.<\/p>\n\n\n\n

If the people we interact with online already think like us, and the algorithm keeps recommending content that fits what we already watch, where do we encounter something that challenges us?<\/p>\n\n\n\n

Maybe the problem isn’t the algorithm itself<\/h2>\n\n\n\n

It is easy to imagine algorithms as something evil sitting somewhere inside a computer, deciding how to manipulate people.<\/p>\n\n\n\n

Reality is probably less dramatic.<\/p>\n\n\n\n

An algorithm doesn’t need to have a political opinion.<\/p>\n\n\n\n

It doesn’t need to hate one group or support another.<\/p>\n\n\n\n

It has a much simpler goal: predict what will keep you interested.<\/p>\n\n\n\n

And what keeps people interested is not always calm or useful information.<\/p>\n\n\n\n

Sometimes it is anger.<\/p>\n\n\n\n

Sometimes it is fear.<\/p>\n\n\n\n

Sometimes it is an argument that makes us want to comment immediately.<\/p>\n\n\n\n

The more we react, the more the system learns.<\/p>\n\n\n\n

And then it gives us more things to react to.<\/p>\n\n\n\n

That is where the cycle becomes interesting.<\/p>\n\n\n\n

We shape the algorithm with our behaviour, and the algorithm shapes the information we encounter.<\/strong><\/p>\n\n\n\n

Neither side is completely independent.<\/p>\n\n\n\n

So, who is really choosing?<\/h2>\n\n\n\n

Perhaps the biggest question is not whether algorithms control us.<\/p>\n\n\n\n

They probably don’t.<\/p>\n\n\n\n

We still choose what to watch. We can unfollow people. We can close the app. We can look for information somewhere else.<\/p>\n\n\n\n

But before we make those choices, something has already made a choice for us:<\/p>\n\n\n\n

What should appear in front of us?<\/strong><\/p>\n\n\n\n

That choice may seem very small.<\/p>\n\n\n\n

One video.<\/p>\n\n\n\n

One post.<\/p>\n\n\n\n

One recommendation.<\/p>\n\n\n\n

One headline.<\/p>\n\n\n\n

But when the same process happens every day to millions of people, it becomes much more important.<\/p>\n\n\n\n

And perhaps that is the uncomfortable part.<\/p>\n\n\n\n

The future of social media may not be about algorithms telling us what to think<\/strong>.<\/p>\n\n\n\n

It may be about algorithms quietly deciding what we get to think about in the first place<\/strong>.<\/p>\n","protected":false},"excerpt":{"rendered":"

What Happens When an Algorithm Decides What You See? Have you ever searched for one video on YouTube and, somehow, ended up watching ten more about exactly the same thing? You watch one video. Then another one appears. Then another. After a while, you start wondering: Why am I suddenly seeing so much of this? The simple answer is that you didn’t choose all of those videos. The algorithm did. Every time we use Instagram, YouTube, TikTok or X, we leave small clues behind. What we watch. What we skip. What we like. What we share. How long we stay on something. The system learns from these actions and uses them to decide what to show us next. It sounds harmless. But what happens when the things an algorithm chooses for us are not just funny videos or football highlights, but news, politics and information about the world? We don’t all see the same internet Imagine two people opening X at the same time. They could live in the same city. They could be the same age. They could even be interested in the same political issue. But their feeds may look completely different. One person might see posts from journalists. Another might see political activists. Someone else might see angry arguments about immigration or crime. The platform is the same. But the version of the platform each person sees can be very different. This is why recommendation algorithms have become an important subject of research. And a recent experiment on X gives us a good reason to pay attention. What happened when researchers changed the feed? In a 2026 study published in Nature, researchers conducted an experiment with 4,965 active X users in the United States. For seven weeks, some users were given X’s algorithmic \u201cFor You\u201d feed, while others used a chronological feed that showed posts in time order. The researchers then compared what happened. The result was surprising. Among people who had been using the chronological feed, switching them to the algorithmic feed shifted some of their political opinions in a more conservative direction. They were also more likely to follow conservative political accounts. The researchers found that the algorithm itself was showing more conservative content than the chronological feed. (Nature) That doesn’t mean the algorithm suddenly turned everyone into a conservative. In fact, the study found no significant change in people’s basic party identification or overall affective political polarisation. That’s an important detail. The researchers found changes in some political attitudes, but not a complete transformation of people’s political identities. And that makes the story more interesting. So, can an algorithm really change our minds? Maybe. But we should be careful. Another large study, published in 2025, gives us a reason not to make dramatic claims. Researchers conducted four experiments involving almost 9,000 people. They created a YouTube-like platform and manipulated the recommendations people received to create different kinds of \u201cfilter bubbles\u201d and \u201crabbit holes\u201d. They found that these changes had limited short-term effects on people’s political opinions. (Harvard Kennedy School) So, if an algorithm shows you more political content from one side, it doesn’t necessarily mean you will immediately change your political beliefs. But perhaps that isn’t the most important question. What if the algorithm doesn’t change your mind? Think about your own social-media feed. If you keep seeing the same kind of content every day, you may not believe everything you see. You may disagree with some of it. But you become familiar with it. You learn the names. You recognise the arguments. You start knowing which issue people are angry about. At the same time, someone else may be seeing something completely different. After months of this, the two of you may sit down and have a political conversation and realise something strange: You are not just disagreeing about the answer. You have been seeing different things all along. And this may be where the real power of recommendation algorithms lies. An algorithm doesn’t necessarily have to tell you what to believe. It may only have to decide what you keep seeing. India is part of this story too This isn’t only an American concern. A 2026 study looked at 825 young Indian voters and examined the relationship between politically similar social networks, algorithmic recommendations and political polarisation. The researchers found that algorithmic curation and politically similar networks were linked with stronger echo chambers and more polarised positions. (ScienceDirect) This was a survey-based study, so it cannot prove that algorithms directly caused the polarisation. But it raises an important question for a country where social media has become a major part of political communication. If the people we interact with online already think like us, and the algorithm keeps recommending content that fits what we already watch, where do we encounter something that challenges us? Maybe the problem isn’t the algorithm itself It is easy to imagine algorithms as something evil sitting somewhere inside a computer, deciding how to manipulate people. Reality is probably less dramatic. An algorithm doesn’t need to have a political opinion. It doesn’t need to hate one group or support another. It has a much simpler goal: predict what will keep you interested. And what keeps people interested is not always calm or useful information. Sometimes it is anger. Sometimes it is fear. Sometimes it is an argument that makes us want to comment immediately. The more we react, the more the system learns. And then it gives us more things to react to. That is where the cycle becomes interesting. We shape the algorithm with our behaviour, and the algorithm shapes the information we encounter. Neither side is completely independent. So, who is really choosing? Perhaps the biggest question is not whether algorithms control us. They probably don’t. We still choose what to watch. We can unfollow people. We can close the app. We can look for information somewhere else. But before we make those choices, something has already made a choice for us: What should appear…<\/p>\n","protected":false},"author":1,"featured_media":436,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-503","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/alexpjose.com\/index.php?rest_route=\/wp\/v2\/posts\/503","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/alexpjose.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/alexpjose.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/alexpjose.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/alexpjose.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=503"}],"version-history":[{"count":1,"href":"https:\/\/alexpjose.com\/index.php?rest_route=\/wp\/v2\/posts\/503\/revisions"}],"predecessor-version":[{"id":504,"href":"https:\/\/alexpjose.com\/index.php?rest_route=\/wp\/v2\/posts\/503\/revisions\/504"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/alexpjose.com\/index.php?rest_route=\/wp\/v2\/media\/436"}],"wp:attachment":[{"href":"https:\/\/alexpjose.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=503"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/alexpjose.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=503"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/alexpjose.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=503"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}