Collagery/ShutterstockAlgorithms are central to modern mathematics and computer science. They are explicit procedures valued for their clarity, efficiency and beauty. Beyond these fields, however, the word “algorithm” is most likely to evoke social media feeds, targeted advertising or a combination of the two of the two.Through algorithms, mathematics has become culturally visible as the force behind opaque processes – with unsettling consequences. Public familiarity with algorithms has grown much faster than familiarity with the mathematical ideas they embody. The changing use of the word algorithm reflects this enhanced visibility. In fact, the meaning of this word has changed dramatically over time, and has always varied between specialist and general use. With the emergence of AI, which is nothing more than a collection of algorithms, this discrepancy has a genuine effect on societal discourse.People sometimes speak of algorithms as they might speak about the weather. Algorithms are simply there, surrounding us, producing broadly similar consequences for everyone while remaining outside individual control. In this sense, algorithms are ambient and impersonal.Yet people also speak of “my algorithm”. For example, the Instagram help pages offer a tool to “tune my algorithm”. Are you seeing lots of news stories about the same event? Are you being shown a lot of adverts for the same product? That’s your algorithm. When people use this language, they are referring to the computations performed on their data, or, more precisely, the recommendations based on them. The possessive is striking. It does not describe ownership. Indeed, few people own, or even understand, the algorithms they refer to. Instead, it expresses a new kind of relationship. The etymology of ‘algorithm’To understand this paradox, it helps to remember that algorithm has never had a fixed meaning. The word pre-dates social media. It pre-dates computers. Indeed, the history of algorithms begins long before there was any machine to run one.The word algorithm developed from a Latinised form of algorism, which refers to a rules-based technique for performing arithmetic. This term can be traced back to the Arabic 9th-century mathematician al-Khwarizmi. The 20th century, particularly through the work of codebreaker Alan Turing and others, gave the word its accepted mathematical meaning: a finite, rigorous procedure, that can be carried out mechanically.Curiously, just as mathematics arrived at their precise notion of algorithm, popular usage began moving in the opposite direction. In contemporary public discourse, algorithm often refers not to a computational procedure but to an enigmatic decision-making system whose internal workings are inaccessible to those affected by it. Where the mathematical expansion was about abstraction, the popular expansion is about relationships. It reflects changing experiences of computation as a source of institutional power.The emergence of expressions such as “my algorithm” marks the latest stage in this history. Within living memory, algorithms were experienced (often unknowingly) as uniform procedures that produced similar outputs for people with a few common characteristics. For example, imagine a bank’s algorithm for calculating loan amounts. The machine-learning algorithms used by social media platforms personalise their behaviour using vast quantities of individual data. The possessive pronoun signals neither ownership nor control, but this personalisation. It names a system that appears uniquely calibrated to a single user while remaining entirely outside that individual’s authority.Seen this way, the contemporary use of the term algorithm is not a corruption of its technical meaning but another chapter in the word’s history. It may be the strangest one yet. For the first time, the word has come to describe something that is simultaneously universal and intimate: a mathematical process that feels deeply personal while remaining institutional and opaque. That paradox says as much about our changing relationship with mathematics as it does about the evolution of a single word.Tom Oliver does not work for, consult, own shares in or receive funding from any company or organisation that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.