No-Training Policy & Data Minimization

Our No-Training Commitment

When analyzing sensitive internal employee communications, confidentiality is paramount. Retaliation Risk Checker enforces a strict Zero AI Training Policy. Any text submitted through our scan engine or API is processed via secure private endpoints. Your drafts are never shared with public model hubs and are never used to train or refine open-source or public large language models (LLMs).

The Principle of Data Minimization

Data minimization is a core tenant of modern data protection regulations (such as GDPR and CCPA). It states that organizations should only collect, store, and process personal data that is absolutely necessary for the intended purpose. In the context of communication compliance, our analysis engine does not need to know the identity of the sender or receiver to evaluate legal risk.

Practical Wording Guidelines

To ensure maximum security, managers should follow these three rules of text minimization before scanning:

• Avoid pasting signature blocks, headers, or email footers containing organizational addresses.
• Focus the scan text strictly on the message body containing instructions, warnings, or schedule discussions.
• Strip out non-essential historical context. Scan only the current message draft, not the entire email chain.