The Taxonomy

A shared vocabulary for describing technology.

Each category is a question people ask about a technology, and its elements are the answers. Choose a category, then an element, to see its icon and what it means. New to the terms? Read the concepts.

What it does

Elements 10

What is this system doing?

This system's actions, said as everyday verbs — it senses, predicts, decides, creates, moves. Naming the verbs makes it easier to reason about what the system touches and what could go wrong.

About this category

What it does

What is this system doing?

This system's actions, said as everyday verbs — it senses, predicts, decides, creates, moves. Naming the verbs makes it easier to reason about what the system touches and what could go wrong.

Datachain

Answer: The purpose often says it already. Name the actions when the system senses, decides or acts in ways the purpose leaves out.
Icon shape: The system itself: why it exists, who is involved, what it does, and the points where it meets the world.

Context: Human involvement

Who completes the loop — a quick, system-level signal of how much runs without a person, shown on the verb it colors. A general characterization of the deployment, not a technical audit.

A person decidesThis system suggests or informs; a person makes the call before anything takes effect. Example: a triage tool ranks cases for a clinician who chooses which to see first.

A person oversees or carries it outThis system works out the result on its own; people supervise it, can step in, or carry out the result themselves. Example: an optimizer plans the day's collection routes, and drivers run them.

Runs on its ownThis system decides and takes effect without a person reviewing each result. Example: a road toll that reprices and charges automatically as traffic changes.

Variables

Description Optional

When to use it

Pick the verbs a reader would recognize; one to three is usually right. To flag autonomy, set Human involvement on the most downstream verb as the system-level signal; set it on other verbs only when their stage genuinely differs; omit it when it would confuse.

Sources

AI Taxonomy — An Operational Framework for Precision in AI DiscourseNarain Jashanmal, v1.1, January 2026. This category extends the source's six modes into a ten-verb vocabulary for public comprehension.

ActsActs — plans a goal into steps and carries them out in software: calling tools, booking, filing, escalating, running multi-step workflows on someone's behalf. The output is an action taken, not a piece of content returned. Examples: booking a meeting, filing a return, escalating a ticket. When the action moves something physical, that is Moves.

AnswersAnswers — responds to people's questions in conversation: a chatbot, a voice assistant, an automated phone line. It holds a back-and-forth addressed to you in the moment. Examples: a transit chatbot explaining a delay, a service line routing your request. Distinct from Creates, which produces content that stands on its own afterward.

CreatesCreates — produces new text, images, audio, video, or code that did not exist before this system ran. The output is content authored by the system, not a label or score about existing data. Examples: drafting a paragraph from a prompt, generating an image from a description, composing audio, completing code. Distinct from Answers, which responds to you in conversation.

DecidesDecides — determines an outcome that takes effect: approving or denying, pricing, allocating, flagging for action. Examples: a dynamic road toll set from traffic, an application approved or refused, a dispatch assignment. When a person makes the final call from its suggestions, that is Recommends instead.

IdentifiesIdentifies — recognizes who or what something is by matching a face, voice, fingerprint, plate, or account against known records. It turns "someone is here" into "this person is here." Examples: a face match at a boarding gate, a plate lookup at a toll gantry. Distinct from Senses, which detects without naming.

MovesMoves — acts on the physical world. The system drives motors, valves, signals, or other actuators and changes something around you. Examples: steering a robot, changing a traffic-signal phase, opening a gate, adjusting a building's ventilation, releasing a parking barrier.

PredictsPredicts — estimates what is likely: a forecast, a risk score, a probability. It says what may happen; someone or something else still decides what to do about it. Examples: a queue-length forecast, a credit-risk score, a crowding prediction for the next hour. Distinct from Decides, which makes an outcome take effect.

RecommendsRecommends — suggests options for a person to choose from: a ranking, a shortlist, a suggested route or treatment. Nothing takes effect until someone accepts the suggestion. Examples: a triage ranking for a clinician, an alternate route offered to a driver, products shown first. Distinct from Decides, where the outcome takes effect on its own.

SensesSenses — turns raw signals from cameras, microphones, or sensors into things the rest of the system can use: detections, transcriptions, readings. Includes continuous monitoring that watches for events. Examples: detecting a vehicle in a camera frame, transcribing a spoken sentence, noticing that a platform is crowded. Distinct from Identifies, which puts a name to who or what it perceives.

UnderstandsUnderstands — pulls meaning from text or speech, finds related ideas, and connects what it reads to saved context. It returns matches, intents, summaries, or links to relevant prior knowledge, not new content. Examples: matching a question to the right help article, retrieving the policy clause that applies to a case, recognizing the topic of a sentence.

Senses

Senses — turns raw signals from cameras, microphones, or sensors into things the rest of the system can use: detections, transcriptions, readings. Includes continuous monitoring that watches for events. Examples: detecting a vehicle in a camera frame, transcribing a spoken sentence, noticing that a platform is crowded. Distinct from Identifies, which puts a name to who or what it perceives.

Context: Human involvement

Who completes the loop — a quick, system-level signal of how much runs without a person, shown on the verb it colors. A general characterization of the deployment, not a technical audit. Choose a value to see its icon.

Variables

Description Optional