Algorithmic curation: Your recommendations choose you – The Cougar

The most effective forms of control do not feel like control at all. They feel like a recommendation. On social media platforms, buttons such as “not interested” and engagement signals appear to grant users control over what content they’d like to see. Algorithms then simplify decision-making by learning what we like. 

But these systems do not merely respond to user preference; they train it, quietly shaping what users come to enjoy, value or believe, while limiting the space for reflection. 

In this way, choice becomes increasingly passive. In this context, how can we claim to be self-curating our feed when so much of what we come across is pre-determined by an algorithm? 

Feedback loops

Degenerative Feedback Loops in Recommender Systems was a 2019 case study that analyzed the feedback echo chamber and filter bubble effects on users. The phenomenon of the feedback loop is the extensive use of machine learning in recommendation systems. 

The decisions made by these systems can influence user beliefs and preferences, which in turn affect the feedback the learning system receives, thus creating a feedback loop.

Feedback loops do not simply reinforce user preference, but they also limit user encounters with opposing views and unprecedented media. On a broader spectrum, curating our feed alters the conditions to which we are naturally curious. 

For example, discovery once arose through accident, curiosity or engagement with outside ideas. Now, algorithmic curation filters the world we engage with by only presenting ideas that are likely to resonate with us. By doing so, beliefs are reinforced before they are even tested, the choice is not fully exercised, it’s shaped in advance by a machine. 

The claim here is not that users are seeing things they don’t want to see, but when choice is anticipated in advance and the chances of encountering the unexpected are designed out, users become a partial product of the algorithm. This raises a deeper question about what kind of freedom is actually being exercised in algorithmically curated environments.

Free-ish 

Political philosopher Isaiah Berlin presents the concept of freedom as unlimited options and contrasts it with the definition of freedom as self-direction. In the first sense of freedom, media platforms appear to facilitate unlimited options through endless content to engage with and the choice to disengage when you want.

Content is recommended to you, pre-filtered, so familiarity is reinforced. Some may argue that the media you consume should be tailored to your liking, and I agree that personalization definitely makes content more engaging and enjoyable. However, the personalization appeal masks how fully voluntary engagement is. 

In other words, tailoring users’ content limits self-direction. Freedom is not choosing among the options that are presented to you, but it requires the ability to encounter the unexpected options. 

Curated algorithms, by recommendations and anticipated preferences, fail to preserve these conditions. This is why feedback loops are inherently dangerous to the human mind because they are trained to suppress the desire and reflection that freedom depends on. 

Antidote to passive choices

Philosophically, freedom requires not just the availability of options but the reflection on the desires that drive those options. In our day-to-day, media-driven lives, we have little space for this type of self-reflection. Users comfortably become the media they consume and justify it on the basis of familiarity or as the product of their own engagement. 

However, when engagement is ranged by design, and curated to your “For You” page, what might have been a conscious decision to engage now becomes a passive choice.

 I encourage users, along with myself, to go on a prolonged passivity with the media, to consume without interacting. The possibility to come across ideas we’ve never thought of, hearing from communities we’ve never encountered, and being inspired by perspectives we’ve never seen would be an ideal antidote to algorithmic curation. 

opinion@thedailycougar.com