AI tokens
Also used: Token
A token is a small unit of text or data that an AI model reads and generates as it processes a request and produces a response.
Why it matters to educators
For educators, token limits affect how much of a curriculum guide, policy, transcript, or conversation an AI tool can consider at once. They can also affect response length, processing time, and cost in some services.
The foundation
A token is not always a full word. Depending on the model, punctuation, parts of words, numbers, and spaces may be represented separately. The tool converts the prompt and response into tokens for processing.
Educators usually do not need to count tokens by hand. The practical lesson is that long inputs have limits: organize source material, include what is relevant, and check whether the tool processed the entire document rather than assuming that upload means understanding.
What this can look like in education
Work with a long handbook
Instead of asking a tool to review an entire handbook at once, a leader separates the relevant sections, identifies the question, and keeps the original document available for verification.
Manage a long planning conversation
After many revisions, a teacher restates the learning objective and fixed requirements so the current response is not shaped only by whichever earlier details still fit in the context window.