No-training data posture
A contractual commitment that a vendor will not use your inputs to train its AI models.
A "no-training" data posture means the AI provider's terms bar it from using your submitted data to train or improve its models. This is the central confidentiality question for lawyers under ABA Formal Opinion 512 and Model Rule 1.6: free consumer chatbot tiers often reserve training rights, while paid API tiers typically do not. No-training is not the same as zero retention: paid providers may still log requests for a limited period, for example for abuse monitoring.
Draftiro does not use your matter content to train AI models. It uses Google's paid Gemini API, whose terms bar Google from using prompts and responses to improve its products; Google may keep requests for a limited period for abuse monitoring. See is AI safe for client data?