Computer Use
Computer use lets an AI system interact with software through a graphical interface. A common implementation uses screenshots for observation and tool actions such as clicks, typing, and scrolling.
[Anthropic]In practice · hypothetical example
An assistant reads a screenshot, selects a search field, types a product code, and checks the next screen.
[Anthropic]A little deeper
The application executes the proposed interface actions and returns observations. Access to a desktop does not establish authorization for every action available on it. [Anthropic]
A common mix-up
Seeing a screen means the model has already clicked it.
Observation and action execution are separate steps. [Anthropic]
What changes the interface after a model requests a click?
Sources & editorial notes
Evidence: supported. Primary-source support for this scoped entry; publication approved by the project owner.
- Computer use tool ↗ (opens in new tab)Anthropic · Publication date unknown
Relevant section: How computer use works; Security considerations
Last editorial review: 2026-09-13 by project-owner.
First observed in this corpus: Unknown.
Revision history
Revision 2 · Created 2026-09-13 · Updated 2026-09-13
Project owner approved the current content for publication. Existing evidence scope and limitations remain applicable.