What we call artificial intelligence
Artificial intelligence brings together methods for tasks such as recognising images, processing language and finding patterns in data. Different systems have different goals and limitations.
Data is part of the outcome
In machine learning, a model finds relationships in examples. Incomplete or unsuitable data can lead to errors in real use. A successful demonstration does not, by itself, establish how a system will perform with other people, languages or situations.
From an idea to a useful tool
First define the task. Then choose measurable criteria: quality, time, cost and the need for human review. Compare the system with the real process you want to improve.
This is an introductory editorial guide, not an announcement of a new product.
