Start with the task
Artificial intelligence is an umbrella term for systems that infer from inputs to produce outputs such as predictions, content, recommendations or decisions. A system that identifies an object in a photo and one that drafts a message do different jobs. Neither needs a moving robot or a chat screen.
A useful way to examine a product is to name its input, its output and what happens to that output. In the authored example below, a photo model proposes a label. A separate application displays it. That label is a prediction to check, rather than direct evidence that the software has human experience of the object.
| Part | Authored photo-app example |
|---|---|
| Input | A photograph |
| Model output | The label “dog” |
| Application action | Show the suggested label to the user |
Conceptual example; no actual model performance is claimed.
Start practicePractice
Apply the idea to
a new situation.
A sound-classification app receives an audio clip and suggests the label “applause.” No source verification or human experience is demonstrated.