The Basics of Recommendation Algorithms

Before we dive into the topic of conspiracy theories and their relation to today’s modern technology, it is important to understand the concept of an algorithm and how it works online. Essentially, an algorithm is a system that has a complex set of rules and instructions that allow it to push, or cover, specific content to the consumer of whatever platform it is working on. An algorithm can help push certain topics; it can also make sure some topics get little notice online. While algorithms follow a set of rules that align with the platform’s motive, their main appeal is how they can push forward content that is relevant and personalized to each user, overall increasing the satisfaction of the users. Social media, online shops, and digital news websites are some of the biggest websites that use recommendation algorithms. While the concept sounds great on paper, algorithms aren’t as perfect as they intend to be, as algorithms today often have a lot of bias and shady practices in the ways they obtain information. Most algorithms track users' online activity on the respective website to get a better understanding of a user, whether that is through tracking what they search, like, buy, share, read, click on, and much more. At first, many were fine with this as it helped push ads and information that was more useful to users and it helped benefit online growth, but as algorithms kept innovating, so did the risk and anti-privacy. Sites began using data to track real-world relationships to increase interest. Algorithms became more corrupt as businesses and political powers influenced the sites to push more of their content to users. However, even with these risks in mind, algorithms have played a vital role in today’s digital media.

Feedback Loop:

Now that we have a deeper understanding of recommendations, we can go into the more complex aspect of the topic, which is the feedback loop. To put it simply, the more a user interacts with content, the better awareness the algorithm has of the preferences and personality of the user. For example, if I were to click on and watch 50 cat videos, the algorithm would most likely start recommending cat-related content and showing ads for things that are feline-related. While logically this process makes sense, the drawback is that often, when a user interacts more and more with one specific topic, the algorithm will start to show more biased and extreme results related to that topic. Taking it back to the cat video example, if I were to watch cat videos, I might start to see only cat-related videos showing up on my feed, and I might see really specific ads marketed toward cat-related items. The algorithm can control a user in this way as they get more sucked into one way of consuming content. This causes a major issue when the bias starts to reflect on the person and their ideas and actions become more radical. This is likely due to a filter bubble, which is when the algorithm limits your worldly knowledge as it only allows you to see a biased form of content. This transitions us into conspiracy theories, as this whole process is one of the reasons for the rise of conspiracy theories.

Conspiracy Theories & Algorithmic Amplification:

The reason why companies utilize the process of the feedback loop is because it increases the engagement time users have with their platform. With the increased engagement time they can push more ads and content on their platform increasing their revenue. However, the increased engagement time allowed for conspiracy theories to flourish as they are able to maintain engagement for long periods of time. Due to this relationship, recommendation algorithms help boost conspiracy theory content as it aligns with its rules and they both benefit from the interaction. But, just how bad are conspiracy theories for society? While I believe conspiracy theories can provide a source of entertainment and serve as a platform where people can challenge the meta social conflicts of society, currently they are being used to spread fear and misguidance. For example, things such as aliens are a fun concept, but when “alien news” is used to overshadow a refugee crisis it becomes a problem. Currently conspiracy theories are used to distract people from what's happening in the world currently, and when major platforms use their algorithms to promote that, it’s an even bigger problem.

Platform Incentives:

With the rising popularity of conspiracy theories, recommendation algorithms have been pushing that type of content more and more. As long as they provide increased engagement, companies will benefit from this cycle. Furthermore, this trend tends to lead to a decrease in the quality of the algorithm as it becomes more broad and convoluted. When algorithms tend to push the content of the conspiracy type, it is harder for them to censor false information and harmful content. As companies are way more focused on revenue, there is a major sacrifice in safety and credibility online. Yet, we as a society allow it to happen as we turn a blind eye to the whole process. If we can’t make credibility and safety a number one priority online, then we can’t expect us as a society to move forward in the development of technology. Companies need to be held more accountable for their actions and not be allowed to overlook the quality of their recommendation algorithms.

Potential for Exploitation and Conclusion:

If we don’t address these issues, here are some potential ways that recommendation algorithms can lead to malicious consequences. Since it's easy to recognize the trends in these types of algorithms, people can purposely alter content or ads to be misleading so that they can be promoted while delivering a different message entirely. Malicious creators can promote content through misleading titles, descriptions, thumbnails, links, and many more mediums to get more engagement. All of these possibilities could run smoothly as the algorithm allows them to slip right through. It is important as a society that we are more educated on recommendation algorithms and that we take a stand to improve their quality and safety. In conclusion, whether you love or hate conspiracy theories, you are responsible for what you push forward online and allow because there is an algorithm tracking every bit.

References

Agrawal, S. K. (2021, July 13). Recommendation system -understanding the basic concepts. Analytics Vidhya. https://www.analyticsvidhya.com/blog/2021/07/recommendation-system-understanding-the-basic-concepts/

Fitzgibbons, L. (2019, April 3). What is Feedback Loop?: Definition from TechTarget. IT Channel. https://www.techtarget.com/searchitchannel/definition/feedback-loop#:~:text=A%20feedback%20loop%20is%20the,first%20stage%2C%20input%20is%20created.

Zeng, J., Schäfer, M. S., & Oliveira, T. M. (2022, August 10). Conspiracy theories in Digital Environments: Moving the Research Field Forward. Convergence (London, England). https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9483695/