Skip to main content
Transparency

About the AI features

A few features on this site use artificial intelligence. Here is exactly what they do, which models run them, who operates those models, and what happens to what you type. No surprises.

Two very different kinds of AI live on this site — and the difference matters.

Retrieval

Most AI here just fetches and organizes what already exists — pointing you to the right provision or scenario. It's an automated table of contents. It doesn't form opinions.

Interpretation

One feature — the questionnaire — is different. It reads your answers and offers a response: a reading of what kind of government those answers describe. That's interpretation, not lookup, and it deserves to be flagged plainly.

When you see AI Feature* with an asterisk, it points here. We'd rather over-explain than have you wonder.

What each feature does, and which model runs it

Feature
What it does
Provision AnnotatorClaude Sonnet 4.6 · Anthropic
Explains the design rationale behind a specific provision, drawn from the constitution's own text. Retrieval and explanation — it doesn't judge you or anyone.
Scenario NavigatorClaude Sonnet 4.6 · Anthropic
Suggests which scenarios and provisions match a question you ask. A recommendation engine — it points, it doesn't opine.
Government QuestionnaireClaude Sonnet 4.6 · Anthropic
Reads your answers to the survey and writes a short response describing the kind of government those answers point toward. This is the interpretive one.

All of these run on Claude, a family of AI models built and operated by Anthropic, a commercial company. When a feature runs, the text it needs is sent to Anthropic's systems, processed, and the response is returned to you.

The questionnaire evaluates your answers — not you

The result at the end of the questionnaire is a response to the answers you selected, nothing more. It is not a judgment of your beliefs, your character, or your politics.

You might pick answers from across the spectrum of political thought, or out of curiosity, or to see what happens — and you'd likely answer differently if you took it again. That's expected. The result reads the answers on the table in that one sitting; it doesn't claim to know you. It's one reasonable interpretation among many, offered to spark thought, not to settle anything.

What happens to what you type

This site's developers do not collect, store, log, or sell your responses. There is no database of answers here, no analytics on what you choose, and no interest in it. Your questionnaire answers exist in your browser for the length of your session and are gone when you close the tab.

However — and this is the honest part — the AI features can't work without sending your input to Anthropic, the company that operates the Claude models. That processing happens under Anthropic's own privacy and usage policies, not ours. We can point you to those policies, but we don't control them, and we can't make promises on Anthropic's behalf. If how your input is handled by the model provider matters to you, please read their policy before using the interpretive feature.

Anthropic's Privacy Policy Anthropic's Usage Terms Contact this site

The short version

The site uses Claude, made by Anthropic, for a few features — mostly to help you find things, and once to offer a reading of your questionnaire answers. We keep none of your data. Anthropic processes what you type under its own policies. And the questionnaire result is a response to your answers, not a verdict on you. That's the whole story.

← Back to the questionnaire