Glossary
What is artificial intelligence
In short
Artificial intelligence (AI) is software that does tasks that used to need human thinking: it understands speech, recognizes images, writes text and makes decisions from data.
AI basics · artificial intelligence
MMOOZZGG.com editorsChecked: 1 min read
For a child
Imagine a toy robot built from a kit that knows nothing at first. Many times you show it pictures of cats and dogs and say which is which. Soon it starts to tell them apart by itself, even in new pictures. This is how artificial intelligence works: a computer looks at lots of examples and learns to do what only a person could do before. It cannot think like you, but it quickly notices what looks alike.
In plain words
AI is the general name for programs that do not follow a rigid instruction but learn from examples. Normal software does only what the developer wrote. An AI system finds patterns by itself: which emails look like spam, which product a customer will like, how to answer a question. When you turn on music by voice or translate a menu with your phone camera on a trip, this technology is at work. Today the word most often means neural networks that write texts and draw pictures.
For business
A company needs AI to take repetitive work off its employees: replies to clients, document review, first checks of applications and demand forecasts. The money comes not from the technology itself but from a specific process where it saves hours or cuts errors. The price depends on the task: a ready cloud service costs less at the start, while your own system on your own servers costs more, but client data does not leave the company. The main risks are wrong answers, leaks of personal data and dependence on one vendor. First find out in which process the investment will pay back, and test it in one area. An AI solution map starts with this kind of review.
For a specialist
Artificial intelligence (AI) is a field of computer science about systems that solve tasks of perception, reasoning, planning and language. In 1950, in the paper “Computing Machinery and Intelligence”, A. Turing proposed to replace the question “can machines think” with the imitation game. The term itself first appeared in the proposal by J. McCarthy, M. Minsky, N. Rochester and C. Shannon for the Dartmouth summer workshop, dated 31 August 1955. Historically there are two approaches: symbolic (rules, logic, expert systems) and statistical, that is machine learning, with deep neural networks inside it. Narrow AI solves one task, and the hypothetical strong AI is called artificial general intelligence (AGI). In the updated OECD definition of 2023, an AI system infers from its input how to generate predictions, content, recommendations or decisions that affect physical or virtual environments.
Example
An online store connects AI to the chat on its website. At night the system answers questions about delivery and returns by itself, and in the morning the managers get only the hard requests.
Related terms
Questions and answers
How is AI different from normal software?
Sources
- A. M. Turing, Computing Machinery and Intelligence, Mind, 1950academic.oup.com
- J. McCarthy, M. L. Minsky, N. Rochester, C. E. Shannon, A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence, 31.08.1955www-formal.stanford.edu
- OECD.AI, Updates to the OECD's definition of an AI system explained, 29.11.2023oecd.ai
Not sure whether your company needs this?
AI solutions map