AI learns patterns from examples

A computer system can be shown many pictures labeled “cat” or “dog.” After finding patterns in those examples, it can make a guess about a new picture. 1 2

AI can make mistakes

An AI system does not understand the world in the same way a person does. Its answer depends on its instructions and examples, so people should check important results. 1 2

Training examples shape an AI system

If examples are missing certain kinds of people, places, or situations, a system may work less well for them. Developers test systems with different examples and check for errors. People should use care when AI results affect important choices. 1 2