Declared Intent
Much of the customer journey has moved inside private AI conversations, and the industry has named the result the intent gap: by the time a person reaches a website, the decision is largely made, and the signals brands relied on have gone dark. The industry's answer so far has been inference, modeling what people have probably decided. This paper argues that asking is the better answer. Inference decays as journeys go private, runs against the direction of privacy law, and gets intent wrong even when all the evidence is there, because the same action can come from very different intentions. Meanwhile the intent the industry is paying to reconstruct already exists: a prompt is a first-person declaration of what a person is trying to do. The paper sets out what legitimate asking requires (specific, revocable consent; compensation; verification that the signal came from a real person; and protection after the data is shared), shows how one signal serves two markets, brands and AI developers, and explains why data that carries the reason behind people's words can be better training data. It argues that declared intent helps people understand each other as well as machines, that intent should be captured on the person's own device with the person keeping the final word, and that every declaration a device model makes should carry a probability so the receiver knows how much to trust it. Humanity Systems working paper. Part of a series on one question: how do we trust AI, now and into the future?
Authors
- A TUCKER
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-14
- DOI
- https://doi.org/10.5281/zenodo.24227102
- Citations
- 2
- Primary Topic
- Privacy, Security, and Data Protection
- Type
- article
- Field-Weighted Citation Impact
- 10.96