What are the privacy concerns when using on-device AI like Gemini Nano?

I think on-device AI is safer. Only problem is if Google still track some data. Company always want data. We cannot fully trust. That is main privacy issue.

Privacy is less concern here, because data not going online. But still some worry: - If phone stolen, data leaked. - If Google update AI secretly, we don’t know what it does. So, we must check permission.