What is artificial intelligence (AI)?

In plain English

Artificial intelligence (AI) is an umbrella term for computer systems that carry out tasks we normally think of as needing human intelligence: spotting a face in a photo, transcribing speech, suggesting the next word you’ll type, or steering a car. “Intelligence” here does not mean consciousness or understanding. It means the system can take in information and produce a useful output without a person writing out every step in advance.

The field is more than seventy years old. Early AI was mostly hand-written rules — “if the temperature is above this number, sound the alarm.” Most AI built today instead learns patterns from large collections of examples, an approach called machine learning.

Why it matters

AI has quietly moved from research labs into everyday infrastructure. It filters your spam, flags unusual card transactions, routes your journey, ranks your social feed and increasingly drafts text and code at work. Because these systems influence real decisions — who gets shortlisted for a job, which scan a radiologist looks at first — questions about accuracy, fairness and accountability are no longer academic.

An example

Your phone’s photo app can show you every picture of your dog, even though nobody ever tagged them. It learned from millions of labelled images what dogs tend to look like, then applied that pattern to your library. That is AI doing one narrow, useful job.

What to watch out for

“AI” is a marketing label stretched across everything from basic statistics to enormous models, so the word alone tells you very little. Ask what the system actually does, what examples it learned from and how often it is wrong. Also beware of assuming breadth: a tool that is superb at one narrow task is often useless, or confidently wrong, at a slightly different one.