By Hoo Jeng Xin

What if the things you think you like are partly things you have been shown enough times to like?

It is an uncomfortable question because we always thought of our preferences as something personal. I like Malaysian food. I like certain kinds of music. I enjoy watching people travel.

I never thought of these things as decisions being made for me. But lately, Australia's social media proposal caught my attention. The federal government has proposed a new “My Feed, My Way” initiative that would give users over 16 a choice between an algorithmically personalised feed and a feed showing content from accounts they have chosen to follow (Australian Government, 2026). The proposal is part of Australia's broader Digital Duty of Care framework, with the government arguing that users should have greater control over what appears in front of them.

At first glance, this sounds like a debate about social media algorithms.

On second thought, I think it is actually a debate about agency.

Because there is a difference between choosing what you want to see and choosing from what you have been given to see.

It led me ponder: how much of what I see online am I actually choosing?

We like to think our feed reflects who we are. Our preferences are being formed in an environment where someone or something is constantly deciding what reaches us. I did not search for Malaysia. Somehow, Malaysia found me. And honestly, I like it.

Every time we stop scrolling, watch something twice, click into a comment section or send a video to a friend, we leave behind a tiny clue about ourselves.

Over time, those clues build up a portrait. But it is not necessarily a portrait of who we are. It is a portrait of what we have done. That distinction matters.

Imagine I spend a week watching videos about moving overseas. My feed might conclude that I am interested in studying abroad, apartment hunting and starting a new life in Melbourne. It may be right.

But perhaps I was only watching those videos because I was anxious about moving.

The system cannot really know the difference. It observes behaviour, not intention.

A 2026 study involving more than 8,800 participants found that people tend to give algorithms narrower versions of their preferences than they give to human curators. Researchers called this “preference filtering”. In simple terms, when we know a machine is trying to understand our interests, we may focus on the things we are most obviously interested in and leave out the less predictable parts of ourselves (Gai et al., 2026).

So the algorithm may not actually know me.

It knows the version of me that keeps clicking.

And then it gives that version back to me.

That is where things get interesting.

We often talk about algorithms as if they are mirrors.

You watch cooking videos, so they show you cooking videos.

You watch football, so they show you football.

You like Malaysian content, so you get more Malaysian content.

Simple.

But a mirror does not change what it reflects, whereas a recommendation system can change what we are exposed to.

A recent study in the Journal of Cultural Economics looked at how engagement-based recommendations can influence the development of people's tastes. The research suggests that when platforms repeatedly control what we are exposed to, they can affect what we become familiar with and, over time, what we come to like (Knight, 2026).

It is more like:

I shape my feed, and then my feed shapes what I notice.

Maybe that is the part we should be paying more attention to.

There is something else we are missing out when everything is personalised: the possibility of stumbling across something we never would have chosen.

Think about how discovery used to work.

You picked up a book because its cover looked interesting. A song playing in a random cafe. You watched a television programme because it happened to be on.

You were exposed to things without first declaring an interest in them.

Now we have almost unlimited choice.

So we built systems to help us choose.

That is convenient. I love that convenience too.

But there is a trade-off.

When a platform decides what appears next, it is not necessarily telling us what to think. It is deciding what gets our attention in the first place.

And research shows that this can create very different online experiences for different people. A 2023 Science study analysing data from 208 million US Facebook users during the 2020 election found strong ideological separation in the political news people were exposed to and engaged with (González-Bailón et al., 2023).

But here is the part that makes the story more complicated.

Another large experiment involving Facebook and Instagram users found that changing from algorithmically ranked feeds to chronological feeds substantially changed what people saw and how they used the platforms, but did not significantly change their political attitudes during the three-month study (Guess et al., 2023).

So no, the story is not simply that algorithms “brainwash” us. They don't.

People still have their own beliefs, experiences and choices.

But algorithms can influence the environment in which those choices are made.

And sometimes, the environment matters.

But what do we actually want from our feed?

Should an algorithm show me what I already like? Or what I might like?
Should it give me what I would most likely click? Or what might actually be good for me?

Those are not always the same thing.

A 2023 experiment with 6,488 participants compared recommendations based on people's actual behaviour based on what people said they ideally wanted. The “ideal preference” recommendations received slightly fewer clicks, but participants reported feeling better about how they spent their time (Khambatta et al., 2023).

That finding changes the question completely.

Perhaps the problem isn't that algorithms know too much about us; the problem is that they are often optimised for the wrong version of us.

The version that clicks, stays, and keeps looking for more similar content.

That is not necessarily the same version of ourselves that wants to learn something, discover something new or close the app feeling satisfied.

So what would “my feed, my way” actually mean?

This is where I struggle with the idea of an algorithm-free feed.

Because I don't actually hate algorithms. Sometimes, I am glad they know me.

My feed has become this strange little space between two countries.

If I completely turned off personalisation, I might gain more control over what I see.

But I might also lose some of the things that make the internet feel personal.

Even Instagram's Adam Mosseri has argued that a purely chronological feed can have its own problems, including being overwhelmed by frequent posts from brands and other accounts. And a recent Guardian experiment found that switching Instagram to a chronological feed made the platform feel less personalised and, after a while, rather dull (The Guardian, 10 September 2026).

So maybe the answer is not to get rid of algorithms completely.

Maybe it is to remember that being shown something is not the same as choosing it.

Sometimes, I think we just need to pause and ask:

Why am I seeing this?
Would I have looked for this myself?
Do I actually care about this?
Or have I just seen it often enough that it feels important?

And maybe most importantly:

What am I not seeing?

Because the biggest thing an algorithm can influence is not necessarily our opinion.

It is our attention.

And attention is how we decide what matters.

That is why Australia's “My Feed, My Way” proposal interests me. It is not just about whether I want a chronological feed or a personalised one.

It is about whether I remember that I have a choice.

My recommendations remind me of Malaysia while I am sitting in Melbourne. It can introduce me to places I would never have searched for. It can even help two very different parts of my life exist on the same screen.

I don't want to lose that.

But I also don't want an algorithm to become the only way I discover the world.

Maybe “my feed, my way” does not mean seeing exactly what I already like.

Maybe it means having enough control to sometimes see something I did not ask for. Something unfamiliar and unexpected.

Because the future version of me should not just be the person my feed predicts I will become.

References:

Australian Government. (8 September, 2026). My Feed, My Way. Prime Minister of Australia. ​https://www.pm.gov.au/media/my-feed-my-way

Gai, P. J., Leung, E., & Klesse, A.-K. (2026). EXPRESS: Preference Filtering: When Customers Share Narrow Preferences with Algorithms. Journal of Marketing. Advance online publication. https://doi.org/10.1177/00222429261466254

Knight, S. Engagement-based curation and the evolution of taste (17 April, 2026). J Cult Econ 50, 399–444. https://doi.org/10.1007/s10824-026-09591-3

González-Bailón, S., Lazer, D., Barberá, P., Zhang, M., Allcott, H., Brown, T., et al. (2023). Asymmetric ideological segregation in exposure to political news on Facebook. Science, 381(6656), 392–398. https://doi.org/10.1126/science.ade7138

Andrew M. Guess et al., How do social media feed algorithms affect attitudes and behavior in an election campaign?.Science381,398-404(2023). DOI:10.1126/science.abp9364

Khambatta, P., Mariadassou, S., Morris, J. et al. Tailoring recommendation algorithms to ideal preferences makes users better off. Sci Rep 13, 9325 (2023). https://doi.org/10.1038/s41598-023-34192-x​

The Guardian. (10 September, 2026). Instagram boss says users will be overwhelmed with brand content in algorithm-free world. https://www.theguardian.com/media/2026/sep/10/instagram-boss-says-users-will-be-overwhelmed-with-brand-content-in-algorithm-free-world