Only algorithms can provide an objective account of one's identity

There are a few different dimensions to our identities. There is the projected self, the perceived self, and the actual self. The projected self, most often expressed in things like clothing, is more who you would like to be, or at least who you would like people to think you are - which obviously does not necessarily say much about who you actually are; it is how you present yourself to the world. The perceived self is how you understand yourself; it is more aware of your insecurities and things you are afraid to present to the world. But it also suffers from the tendency we all have to be biased in our own favor. The actual self is a matter of objective record. But it is virtually impossible to discern, as the mind is not like some lab rat we can observe and dissect (which is my main gripe with the social sciences, at least to the extent that they forget they dwell more in the philosophical realm to which we can make imperfect attempts to introduce scientific rigor; i think they often view themselves the other way around: as scientifically rigorous things from which philosophical truths can be derived) and, like the perceived self, is hard to see because of the bias we all have in our own favor, and desire to adhere to the narrative we have each constructed about ourselves.

However, if an algorithm is something that observes your behavior, learns how you think, what you like, what you dislike, and then shows you things that conform to this, is it not the closest possible thing we have to an objective record of the self? Theoretically, if I were to predict with 100% accuracy what, in a complex jungle of information and opportunities for action (i.e. someone is faced with lots of possibilities and must choose one), someone would gravitate to, could it not be reasonably said that I have an objectively valid understanding of who that person is? Is someone not essentially just what they do when given choice? In the way we understand identity, I don’t think this is perhaps entirely accurate. It is probably more downstream from the actual self; but it can be produced by little that is external to the actual self. If we could put everyone in a maze of unlimited opportunity, with the range of opportunities following an initial action specific to that initial action (i.e. path dependency), I am confident that no two people who navigate this maze of opportunity in the same way.

Naturally, these algorithms do not always get it right; but given the near unlimited amount someone can be shown on Tik Tok or instagram, that the algorithm can guess with even reasonable accuracy what might drive engagement, this would suggest to me that at the very least they understand us far more accurately than we understand ourselves. Indeed, are we each not essentially the sum of what we like, what we don’t like, what we find interesting, what makes us sad, what makes us happy? And are the decisions (likes, saves, views) we make when uninhibited not a near-perfect manifestation of these things? Algorithms are unique in that they are an understanding of the self that does not require taking an individual on their own terms. They do not dwell in the otherwise contextual and subjective elements of objectivity, they simply know what we will like. Even when the things we like contradict what the person we perceive ourselves to be would like. And on the surface this would perhaps seem largley irrelivent to the extent that people are largely just following shit-posting meme accounts, because they are just jokes and there is nothing that says much about who we are by liking them. But we dramatically underestimate the substance of a meme. Memes generally convey something that has happened and then take a view of the thing that has happened. And we tend to like the memes who take a view of something that we share. Gone are the days of memes that were essentially just one dimensional captions on stock images. Because we live in a world where current events, even if they are events that do not seem worthy of a news article, are made into memes almost as soon as they occur, things that happen in our world are instantly conveyed, as are the perspective taken by each meme. Thus, liking a meme on instagram actually conveys a lot more than simply finding something funny; because memes generally articulate a view of something, it conveys which views you share.

It would be interesting if there was a way to somehow reverse engineer, say, someone’s tik tok algorithm and have it spit out the algorithm’s understanding of each of us. I think the reason we, for good reason, regard identity as a very nebulous, personal, undefinable, and subjective (in the sense that it is not a matter of record; there is probably a better word than subjective here) thing. But that is because before the current times there was nothing that really effectively captured, tracked, and responded to how we navigate a large, complex maze of stimuli.

Even taking this a step further, what would happen if we applied algorithms to daily life? What if your tik tok algorithm could be used to predict how well you would do at a job you’d apply for? What if it were used to predict whether you’re likely to commit a crime?

We often think we have a perfect understanding of those we are close with. But in reality we each have recesses of our minds that we keep from others. But there is something freeing about being on social media. We peruse it in an unguarded manner. Our actions on social media do not have to conform to our sense of ourselves, they do not have to be aspects of ourselves that we have to be able to articulate to the world in a way the world would like and understand, and how we would like to present ourselves to the world. They are who we are when we are entirely uninhibited.