What is Artificial Intelligence (AI)?
Artificial intelligence (AI) is software that performs tasks which normally need human judgement, such as recognising images or speech, understanding language, predicting an outcome, making a decision or generating text and pictures. Most AI today learns how to do one task from large amounts of example data rather than following rules written by hand.
The term covers a wide range of systems, from a spam filter that sorts email to a model that writes code. What they share is that the behaviour comes from patterns found in data, which is why the quality and quantity of that data matter so much to the result.
How does AI work?
- Collect examplesA system is built from a large set of examples of the task, such as labelled photos, past transactions or pages of text.
- Train a modelAn algorithm works through the examples and adjusts millions of internal settings until its answers match the examples as closely as possible. The result is a model.
- Make predictionsThe model is given new input it has never seen and produces an answer: a label, a number, a ranking, a sentence or an image.
- Measure and updateIts answers are checked against real outcomes. When accuracy drops, the model is retrained on newer data, because the world it learned from keeps changing.
What are examples of AI?
- RecognitionFraud checks on a paymentA bank's model has seen millions of transactions, legitimate and fraudulent. When a new card payment arrives it scores the risk in milliseconds and flags the unusual ones for review.
- PredictionRecommendations in a streaming appThe app records what each person watches and compares it with the habits of millions of others, then predicts which titles that person is most likely to start next.
- GenerationAn assistant that drafts repliesA language model trained on a very large body of text reads a customer's message and writes a reply in the right tone, which a person checks before it is sent.
Which terms are related to AI?
- Machine learning (ML)A way of building AI where a model learns patterns from data instead of following rules written by hand.
- Generative AIAI that creates new text, images, code or audio from a prompt rather than classifying existing data.
- Artificial general intelligence (AGI)A hypothetical AI that could learn and reason across any task at a human level, as opposed to today's narrow systems.
- Deep learningMachine learning with many-layered neural networks, the technique behind image recognition, speech and large language models.
Common questions about AI
Is AI the same as machine learning?
No. Machine learning is one way to build AI, and today it is the most common way. AI is the broader goal of software that performs tasks needing human judgement; machine learning is the method where a model learns those tasks from data. Rule-based expert systems are AI without machine learning.
What is AI used for in everyday products?
Spam filtering, search ranking, product and content recommendations, fraud checks on payments, speech-to-text, photo organisation, route planning, customer service assistants and drafting text. In most of these the AI runs in the background and the user only sees the result.
