What are algorithms?
Algorithms. If it weren’t for COVID, they’d be claiming “most disliked” awards in the end of year round-ups. Surprising levels of dislike, considering they are around us daily, our entire lives. In recent times they’ve had particular focus around social media algorithms, and in academic rankings. If it’s got to that stage where it’s got awkward to admit you aren’t exactly sure what they are, here’s our handy introduction…
What is an algorithm?
Definition: “a process or set of rules to be followed in calculations or other problem-solving operations, especially by a computer”. Basically it’s how you decide how to sort information into an actionable series. Whether that’s what social media posts to show a user first, or who receives the highest grades, or perhaps most currently, which people are given a vaccine first… An algorithm is functionally uncontentious (life can’t exist without them) but morally contentious, as every algorithm is the product of decisions by its creators… and therefore their intentions, background and biases.
When did they first appear?
Well, they’ve always been present, ever since mankind recorded information. Arguably before that, if one includes the series of considerations our ancestors took over who ate the kill from the hunt first (weakest? Strongest? Highest stats?). A Persian dude in the 9th century gets credit. In the twentieth century, the concept had a fresh resurgence, beginning with a German gent.
As ever, Wikipedia has a banging timeline of algorithms. Don’t tell us we don’t know to party. We know how to hook you up.
When did algorithms start to matter so much?
Well, without getting into social-cultural political theory, we’re going to point fingers at Ada Lovelace. Quite a badass herself, she broadly gave us computing machines (or the thing you’re reading this on). Machines getting involved in algorithms has really changed the game.
Some of it is just fear: why should a machine be more alarming than a person when it makes decisions? Arguably they’re more reliable.
Other factors are more valid. We have vastly more information, and can sort it vastly faster. A newspaper editor used to decide between hundreds of potential stories for your edification – your social networks / Google now do the same but across billions of potential ‘stories’, and not even just daily – every time you refresh. So, there are more opportunities to influence, to share, to distort.
Globalisation means algorithms are often determined by those outside your legal jurisdictions. Whether that’s for better or worse depends on both where you reside, but also your perspective. None the less, it is less predictable and controlled than in yore.
Possibly most long-term significant is that machines (computers, for the layperson) are now able to self-train, meaning that whilst we set them off with the parameters of an algorithm to solve a particular problem, they then evolve based on what they find, using self-referencing instructions. After a while, no single person knows the algorithm. Facebook has hit this point. i.e. the algorithm is bigger than any single set of instructions.
The fact you’re reading this at all is the consequence of multiple algorithms. Some, distantly, involve the basic fact the author has received certain educational and life opportunities, determined by the existence of algorithms, allowing them to write this content. More directly, they include the basic filtering ot the inbox/social feed/Google search which led you here, based on a complex series of algorithmic assessments of the most relevant content to meet your needs.
Algorithms are themselves neutral. It’s our framing of them, management of them, and judgement of them which is subjective. There are as many algorithms as there are opinions. None of them are right – they just sit in a spectrum of ‘rightness’, based upon your personal framework.
Give them a break. We need them.