The Tyranny of Recommendation Algorithms and Collective Extremism

Recommendation system algorithms are a technological advancement widely adopted by major online platforms, shaping much of what we see online today. However, many people remain unaware of the broader implications of these systems.

Originally, recommendation systems were designed to address the problem of information overload. As the volume of content—books, newspapers, and digital media—grows exponentially, individual consumption struggles to keep pace. Even popular forums generate so many new posts daily that managing them often requires numerous moderators. In just a few decades, humanity has transitioned from an era of information scarcity to one of information overload. To navigate this vast sea of content, innovative technological solutions have become essential.

For this reason, recommendation systems are introduced. A fundamental principle of recommendation systems is “like attracts like”, both in terms of items and people, and this principle is embodied in algorithms. For example, if someone enjoys reading J.R.R. Tolkien’s Lord of the Rings trilogy but is unaware of the existence of the Fall of Númenor, the platform might use algorithms to automatically recommend this and other Middle-Earth writings. Many people would likely welcome such recommendations. If the recommendation system identifies someone as a fantasy fiction fan, it will suggest books of such genre, instead of recommending something like Chicken Soup for the Soul.

A significant drawback of recommendation systems is their tendency to exacerbate societal biases rather than mitigate them.

People generally tend to engage with content that aligns with their existing preferences and beliefs, and this means algorithms will reinforce these preferences, effectively cementing narrow and biased perspectives. This can be likened to confining oneself in a mental bubble where the perceived scope of the world is limited to this narrow circle.

For instance, individuals with a particular viewpoint may find that 85% of their social media news feed aligns with their opinions, thus further reinforcing their beliefs. Conversely, those with opposing views might see 90.85% of content supporting their stance, which can further entrench their beliefs. This skewed exposure can lead to a distorted sense of consensus and a perception that opposing views are far less prevalent than they actually are. Consequently, both groups may believe they are in the right, despite the skewed nature of their information landscape.

This is how observational errors and biases are generated. Recommendation algorithms craft a distorted sense of reality by curating what we see based on our preferences, creating a narrow, personalized information bubble. It is crucial not to confuse this limited perspective with the broader reality. The information bias produced by these systems, among other factors, can easily lead to a form of cognitive reinforcement that shapes our lives and identities.

As a consequence, society becomes more polarized and extreme. This phenomenon is closely linked to the influence of recommendation algorithms, which distort perceptions by reinforcing biased viewpoints.

Unsurprisingly, in today’s world, many people tend to focus on a single viewpoint at a time. Amid the tyranny of recommendation algorithms, suppressing opposing perspectives is frequently seen as a form of justice. A nuanced and objective understanding often rests with a select group of specialists, who can provide a range of diverse and in-depth information. However, due to issues like information overload, observational bias, and content filtering, even these experts may find it challenging to gain wider recognition.

Have you read?
How Senior Leaders Can Elevate Employee Performance and Accelerate Retention.
R U OK Day: How CEOs Can Build Emotional Resilience in Their Teams.
The 3 Secret Ingredients for Making Big Bet Decisions.
Building a culture of commitment.
Corporate Social Responsibility (CSR): A CEO’s Perspective on Integrating ESG for Sustainable Growth.