AI Policy

Last updated:

Which ArcDemo features use AI, what data they send, and our commitment never to train on your content.

The short version

  • AI features run only when someone in your workspace uses them.
  • Neither we nor our AI providers train models on your content.
  • Always review AI output before you publish it.
On this page
  1. Where ArcDemo uses AI
  2. What data is sent
  3. No training on your content
  4. Keep a human in the loop
  5. Transparency for viewers
  6. Prohibited uses

Where ArcDemo uses AI

  • Editor agent: rewrites copy, fixes steps and explains changes when you ask it to.
  • Translation: creates language versions of a demo.
  • Voiceover: turns step scripts into narration, and transcribes recorded narration.
  • Buyer agents: answer viewer questions using the demos and documents you provide.

What data is sent

Only what the task needs, such as a step's text, a screenshot, or a voiceover script. It goes to the providers listed on our Subprocessors page, over encrypted connections.

No training on your content

We don't use your content to train AI models. We use our providers under business API terms that don't let them train on it either.

Keep a human in the loop

AI can be wrong or incomplete. The editor agent shows you each change before it's applied, and changes can be undone. You're responsible for what you publish.

Transparency for viewers

Buyer agents tell viewers they're talking to an AI. If you publish synthetic voices or AI-generated media, label them where the law requires it, including under the EU AI Act.

Prohibited uses

Don't use ArcDemo's AI features to impersonate real people, clone a voice without the person's consent, or create content that breaks our Acceptable Use Policy.