Every second of every day, decisions are being made on your behalf by systems you never agreed to use and can't fully see. The price you're quoted for a flight. The job listing that surfaces at the top of your search. The song that plays next. The friend whose post you notice — and the one whose post disappears entirely. These outcomes are not random, and they are not curated by humans sitting at desks. They are the product of algorithms: mathematical procedures designed to take inputs, apply rules, and produce outputs at a scale no human team could match.
The word "algorithm" gets used loosely, often to mean something vague and powerful and slightly sinister. The reality is more specific. An algorithm is simply a defined set of steps for solving a problem. But the algorithms that govern digital life today are not simple. They are trained on billions of data points, optimized for measurable goals, and applied to populations of millions. The gap between the goal an algorithm is designed to optimize for and the outcomes it actually produces — in human terms, in social terms — is where most of the controversy lives.
Some of these algorithms are decades old, developed in academic settings long before the internet made them commercially essential. Others were designed specifically for the attention economy, built to keep eyes on screens and clicks flowing. A few are so woven into the infrastructure of the internet that removing them would require rebuilding systems from scratch. All of them make trade-offs. Speed versus accuracy. Personalization versus privacy. Efficiency versus fairness.
