Privacy controls for sponsored GenAI
Using an AI chat involves distinct data flows: the content needed to answer your request, local browser state and information used to display or measure an ad. These flows have different purposes and controls.
How do privacy controls work in sponsored GenAI?
Your prompt and any content needed for the answer follow the model-provider path. An ad runs separately and does not receive permission to change that request or its result.
The chat's privacy policy explains provider processing and browser storage. Optional analytics and semantic targeting have their own consent conditions; accepting one does not authorize every use of your data.
What data rules must stay explicit?
A responsible sponsored GenAI product makes prompt handling, provider sharing, and optional telemetry easy to inspect. The product should not ask users to infer these rules from vague UI text.
- • What the model provider receives
- • What remains local to the app or browser
- • Which optional categories require consent
- • What the ad market never receives
What does the ad market receive?
Prompts and chat history are not sent to advertising partners. In a partner-connected integration, wavebird reduces permitted context to delivery signals such as broad category and language, alongside technical placement information.
This public chat uses wavebird's own house placements to demonstrate the experience. Seeing a placement here does not mean your prompt was sent to an external advertising network.
How it connects back to wavebird.ai
For your own application, review the data firewall and consent documentation on wavebird.ai. They explain which fields your backend sends, what wavebird processes and which reduced signals can reach an advertising partner.
FAQ
Does the ad market receive prompt text?
No. Prompt handling and provider sharing are separate from the ad-market boundary. The product should make that distinction visible and enforce it technically.
Are privacy controls only a legal page topic?
No. They belong in settings, in-product disclosure, and architecture design, not only in the privacy policy.
What is the relationship to consent?
Consent governs optional categories such as non-essential measurement or stronger personalization. Necessary provider processing for answering the request is a separate product boundary.
Where are the deeper product definitions?
The data firewall and API documentation on wavebird.ai describe the integration. This chat has its own privacy policy for the service you are using here.