AI REGULATORY SANDBOXES / INDIVIDUAL MAPPING
THE WORKSHOP PULSE

Where do we
draw the lines?

Explore what makes a sandbox, what it is for,
and where other experimentation environments overlap.

● SIMULATED DATA

30 fictional viewpoints.
Designed for rehearsal, not evidence.

Q2 · ESSENTIAL CHARACTERISTICS

What makes it a sandbox?

Up to 3 choices

Read “Other” responses
Q3 · TYPOLOGICAL DIMENSIONS

How do we distinguish types?

Up to 3 choices

Read “Other” responses
Q4 · OVERLAPPING ENVIRONMENTS

Shared ground. Different boundaries.

One rating per environment. Bars show the distribution among those who answered each row.

Strong overlapSome overlapLittle / no overlap
Other environments mentioned
Q1 · GOALS

What should a sandbox achieve?

Theme labels are supplied annotations, not automatic analysis. Select a theme to read its source responses.

Q5 · THE DEFINING DISTINCTION

Beyond a label.

Different answers to the same boundary question.

DATA & METHODOLOGY

Explore patterns.
Do not manufacture consensus.

The demo contains 30 individually authored fictional responses: six professional perspectives, five people per perspective, and 15 participants in each modality. No actual participant, organisation or survey response is represented. This is a deliberately varied rehearsal dataset, not a representative sample or a prediction.

What is measured?

Q1 asks for goals; Q2 and Q3 allow up to three selections; Q4 rates four environments (plus optional Other); Q5 asks for a one-sentence distinction. Selection percentages use respondents answering that question and may sum above 100%. Missing answers are excluded from that question’s denominator.

How are open answers handled?

Q1 and Q5 theme tags were manually assigned for the simulation. Connections count respondents with both tags and both open answers. They may overlap. Imported data is not automatically coded or summarised. Every theme can be traced to its original text. No automatic “consensus” threshold is applied: use the workshop discussion to decide whether a claim is consensus, contested or unresolved.

What is additional metadata?

Profile, persona and modality are simulation metadata, not extra questions in your five-question survey. To use these filters with real responses, collect the needed metadata separately and voluntarily. Missing metadata is shown as Unspecified. A professional profile does not imply a particular position.

Using your own data

Import replaces the demo in this browser tab. The imported file is processed locally, is not saved to the Site, and disappears on reload. The Site does not collect survey responses or connect to Google Forms. Export the data you want to keep. Remove names and identifying details before presenting responses.

Download demo CSVDownload demo JSON
 BRING YOUR OWN RESPONSES

Replace the demo.

Choose a CSV using the template below, or a JSON array matching the demo file. Data stays in this browser tab; it is not uploaded or stored. This is a file import, not a live survey connection.

Remove names and identifying free text before use. Profile and modality are optional. Blank answers are allowed and excluded from their question’s denominator.

Download CSV templateExample JSON ↗
Columns, codes and mapping instructions