How AI Works, Without the Maths

An alarm clock can ring at 7 am without learning anything about your morning. Someone programmed a rule: when the clock reaches the chosen time, sound the alarm. Plenty of useful software works through rules and calculations like this. Now imagine helping a community club run a creative-writing workshop. Messages arrive saying “Can I come along?”, “Please reserve a place” and “Is there room for me?” A rule that looks only for the word “reserve” would miss two of these. A system trained on examples of booking enquiries could learn to recognise different ways of expressing the same request. This is machine learning: adjusting a model using examples so it can handle new inputs. A model that assigns a category is often called a classifier. A generative model produces content, such as a reply to a booking enquiry. A language model learns patterns in text that help it produce language, including explanations, summaries and invitations. These operations can work together. A learned classifier could identify a booking enquiry, a rule could send it to the organiser, and a language model could draft a reply. The surrounding application connects those steps. AI is a broad field; a chatbot is one kind of AI application, and automation does not always require AI.