use private AI for sensitive topics

Pattern

Sensitive context belongs only in AI systems with privacy boundaries that match the risk.

Use private, preferably local AI for sensitive topics, or remove the sensitive details before involving a remote model.

For one sensitive task, use a privacy-first AI assistant or a model that runs on your own device.

Local execution still depends on the security of the model, interface, logs, extensions, and host system.

local-first: Sensitive or durable information stays under personal control before remote convenience is added.