Responsible AI Principles & Policy
At DhiTantra, AI is designed to augment human legal judgment, never replace it. Our engineering and deployment protocols are strictly governed by the following enterprise principles.
What this means in practice
- AI assists. Important answers should be reviewed by the user.
- Legal citations are verified against trusted legal materials.
- Confidence varies depending on the legal information available for a given question.
- Users remain responsible for the legal decisions they make using DhiTantra.
Our Principles
Transparency
We are clear about what DhiTantra is, what it draws from, and where it can fall short.
Explainability
Outputs are linked to their source materials so reasoning can be reviewed and challenged.
Privacy
Workspace content stays inside your workspace. It is not used to train shared models without explicit consent.
Human Oversight
DhiTantra assists professional judgment - it does not substitute for it. Important answers should always be reviewed.
Continuous Improvement
We invest in better legal grounding, evaluations, and feedback loops so the platform improves over time.
Known Limitations
Generative AI systems may occasionally produce incorrect, incomplete, or outdated information. DhiTantra is designed to assist professional judgment, never to replace professional responsibility. All generated outputs must be reviewed by qualified personnel before legal reliance.