AI Use Policy
This policy explains how Trooth uses artificial intelligence in our product and internally, the guardrails we keep around it, and how you can raise concerns. It applies to everyone who builds, operates, or relies on Trooth.
How Trooth uses AI
Trooth uses AI to help you move faster, never to make a final call on your behalf. The two main places you will meet it are:
- Kyrie, our assistant, which answers questions about your posture and helps you navigate the product and frameworks.
- Questionnaire drafting, which proposes answers to security and vendor questionnaires from evidence Trooth has already witnessed, so you start from a draft instead of a blank page.
Every AI-assisted output in the product is labelled as AI-generated at the point you see it, so it is always clear when you are reading a draft rather than a witnessed fact.
Human oversight
High-impact outputs are reviewed by a person; you keep the pen. Trooth drafts, suggests, and organizes, but a human decides what to submit, publish, or rely on. We design AI features so that a person is always in a position to review, edit, or reject the output before it has any real effect.
Trooth automates. Trooth never signs. The platform will not certify, approve, or attest anything on your behalf.
Training-data stance
Trooth does not use your data to train AI models. Your content, evidence, and connected-integration data are used to provide the service to you, not to train foundation models, and we do not sell or share your data for others to train on. This is the same promise you see inside the dashboard, stated here in public.
Reporting misuse
If you believe an AI feature has behaved unsafely, produced harmful output, or been misused, tell us at hello@trooth.co. We review every report and, where appropriate, adjust the feature or its guardrails.
For the specific models, data access, and limitations behind these features, see our model card.