Algorithm

An algorithm is a precise set of instructions for solving a problem. Think of it as a recipe: it takes inputs, follows a defined sequence of steps, and produces an output. Every piece of software you use runs on algorithms, and AI is no exception.

Algorithms in everyday life

You encounter algorithms constantly. A search engine uses algorithms to rank web pages. Your GPS uses algorithms to find the fastest route. Social media feeds use algorithms to decide which posts you see first. The word has become associated with mystery and manipulation, but at its core an algorithm is just a procedure.

Algorithms in AI and machine learning

In machine learning, algorithms are the methods a model uses to learn from data. Common examples include:

• Linear regression — finds the best straight line through data points to make predictions
• Decision trees — splits data into branches based on feature values to classify or predict
• Gradient descent — iteratively adjusts model parameters to minimise errors during training
• Backpropagation — calculates how to update each weight in a neural network to reduce mistakes

The choice of algorithm affects everything: how fast a model trains, how accurately it performs, what kind of data it needs, and how interpretable its decisions are.

Algorithm vs. model

An algorithm is the process. A model is the result. When you train a neural network using the backpropagation algorithm on a dataset of customer reviews, the trained network — with all its learned weights — is the model. The algorithm is how it got there.

Why this matters for non-technical people

When someone says "our AI algorithm does X," they are describing the method their system uses. Understanding that algorithms are chosen, not inevitable, helps you ask better questions: Why this algorithm? What are the trade-offs? What assumptions does it make about the data?