I wrote this blog article first for Knowledge@emlyon.com, based on research published with Massimo Airoldi in 2022 :
In the same way that algorithms influence our everyday experiences, we have the ability to influence the algorithms. But how exactly? Algorithms are here to stay and they are part of our culture from now on. We therefore need to try to understand how they work.
We live in an algorithmic age where much of what we experience is controlled and shaped by digital platforms and their algorithms. Yet, given that algorithms are owned by companies, we know only very little about them and often we are not even aware of them. As scholars, it is hard not to ignore this problematic issue. We are facing something that may easily foster utopian or dystopian illusions and thoughts.
On the one hand, the new technology offers many opportunities for people and companies. We may have a chance to receive recommended content that interests us. Spotify spurs tips on music and artists that we did not know of but – based on our prior listening behavior – would likely interest us. Netflix finds movies and TV series in the same way, and our Instagram feed is populated by content optimized by the algorithm’s calculation, in ways that takes into account the content we are following and interacting with. For scholars, platforms such as Google Scholar or Research Gate recommend research articles that we should be reading. We can just sit back and enjoy, while the algorithm is doing the job for us?
On the other hand, there is a potential dark side to companies’ endless data collection, surveillance and exploitation practices. They can not only tell what we have been doing in the past but predict which choices we are likely to make in the future, too – also shaping us in ways that align us with their purposes and ideology of consumption. A consequence can be that the algorithms and powerful machine learning techniques lead to greater corporate manipulation.
In our recent research article, published in the Consumption Markets and Culture, we examined perspectives toward understanding this cultural shift. Instead of painting a picture of either a utopian or dystopian condition, we hoped to nuance a more holistic understanding of how algorithms influence us. We highlight that companies’ attempts to control, or “nudge”, people also depend on their more or less active resistance.
Algorithmic consumer culture is characterized by four features. They include:
- Opaque. Hardly anyone, even the experts themselves, has an insight into how the algorithms are structured and how they do calculations and classifications based on data. They are owned by private companies and thus are part of their commercial secrets. Also they are driven by AI solutions that constantly feed in new input to get more accurate output, and they are therefore constantly changing, explaining why it is even more difficult to keep up with how they function.
- Authoritative. They want to direct us toward certain choices. All marketing communications want to control us in the same sense – buy three, pay for two! – but this is perhaps another level of influencing. The algorithms create our entire digital world, making us perceive and relate to the world in a specific way, and to act according to what is reasonable given this standpoint.
- Non-neutral. Algorithmic systems depend on cultural assumptions in their construction. If you like this, then you probably like this too, if you have this profile, then chances are you like similar things to others with similar kinds of profiles. Such assumptions are fed into the mathematical models behind the algorithm and because the systems constantly retrieve new input data as a feedback from people, then such assumptions get reinforced and emphasized in the future content recommendations. Eventually, as other scholars have shown, this can facilitate the polarization of stereotypes, which guide our actions, and the content that is made visible (or invisible) to us.
- Recursive. The algorithmic systems are based on user data being constantly fed into the systems, which with the help of machine learning are constantly tweaked to function even better. The systems therefore control us humans, but we also control the systems themselves. This recursive relationship is what makes the systems work – and thus our chances for intervention and resistance. Yet, should we not pay critical attention, they may keep getting more extreme, in the sense of corporate control and manipulation, making us act according to how the companies want.
In short, this is what the algorithmic systems look like. What does this then mean for us, having to navigate the consumer culture increasingly co-created by the algorithms? It appears there is a constant negotiation between control and resistance that we have to acknowledge. On the other hand, we are controlled by the algorithms, but we also control the algorithms in turn, and this can influence the way they end up developing. We can see this dynamic on three levels:
- Individual level. Well-adapted algorithms can assume the guise of “technological unconscious” (Beer 2009) that controls and mediates our desires in ways that we do not fully comprehend. Without really knowing how it happened, we surfed away online, clicked on something, and we maybe even feel good about it. A world is created in which we feel natural. But maybe we can also sometimes be disgusted about the fact that we are running into content that we did not choose ourselves. There are individual ways to resist such as blocking advertising, refusing cookies, refusing to release data between apps, or simply providing incorrect information about ourselves when given the opportunity. As an individual, you have understood that all data is collected and used and optimized for you based on a commercial logic.
- Collective level. The social exchange between different individuals is tightly controlled in social media. Most of us have experienced how we only see a fraction of our “friends” on Facebook, Instagram, TikTok or whatever platform we prefer. These “friends” are also ones with whom we share a lot in common, so that filter bubbles are created based on what everyone seemingly agrees on. It looks like this because companies have figured out that this makes us spend more time on their platforms, and more time means more data and money for them. However, there are several examples of how groups of consumers can mobilize and even hijack the algorithmic logic for their own purposes. This is what we can call “algorithmic resistance”. A poignant example is when Reddit users banded together and managed to drive up GameStop stock through a resistant collective action.
- Market level. One of the domains in which market actors use algorithmic systems to influence and control is to classify and rank different cultural manifestations. This is manifested not least in the search engines’ constant attempts to offer us things we appreciate when we search for a specific concept. These attempts to predict what we will appreciate also lead to simplifications and, not least, to the reinforcement of stereotypes. This has been criticized in various ways and there also exists ways to resist. An interesting example is Johanna Burai’s internationally recognized art project “World-White-Web”. The idea was to influence the search algorithm in a way that it would not only return images of “white people”, and instead began to represent different ethnicities and skin tones, when searching for the word “hands”. A small resistance, perhaps, but nevertheless an illustration that it is possible to change the system on its own terms.
In conclusion, we don’t agree that algorithms are bound to creating either utopian or dystopian worlds. It’s easy to keep with old habits and routines, shrug one’s shoulders and blame the algorithmic culture. But algorithms are here to stay and belong to the culture from now on. It is our task and responsibility to be reflexive about how we interact with them, and figuring out how they influence us. Culture is not something external out there, but it exists because we reproduce it every day. Algorithms require input from people in order to function, and thus they largely reflect and re-iterate patterns of our behavior. Thus they do not control us more than we control them.
[This article is based on the academic publication: Airoldi, M. and Rokka, J. (2022) “Algorithmic consumer culture”, Consumption Markets & Culture, 25(5), 411-428. https://doi.org/10.1080/10253866.2022.2084726.]
