Responsible AI

Approach

XRDNA's approach distinguishes information retrieval, recommendations and human-authorized action, with access constrained by the permissions and operating context of the deployment.

Limitations

AI output can be incomplete or wrong, and needs appropriate validation — especially in consequential mission or safety contexts.

Documented per implementation

Whether a feature uses a third-party model service, retains prompts, connects to sensitive data or can initiate actions is documented for each implementation.