ReelPeel: Inquiry under Algorithmic Feeds via Hold-to-Enter In-Feed Evidence Inspection
Year: 2026
Type: conference-paper
Abstract: Short-form video is increasingly used for health information, but verification remains costly because claims are transient, often spoken, and embedded in a fast scrolling feed [2, 3, 12]. ReelPeel is a browser overlay for Instagram Reels that turns this verification step into an in-feed interaction. Users enter an Inquiry Mode via press-and-hold, inspect extracted medical claims, and progressively drill down into ranked evidence cards with provenance, publication type, relevance, stance cues, and direct source links. Rather than collapsing the analysis into a single credibility label, ReelPeel supports claim-level inspection under viewing-time constraints. The system combines a browser extension with a configuration-driven analysis backend that transforms a Reel into structured claim and evidence objects. In the current medical prototype, this includes PubMed retrieval and optional retrieval-augmented guideline evidence. ReelPeel contributes an AI-enabled interaction and an extensible architecture for inspectable evidence access in short-form video.
Language: English
Subfield: Sociology and Political Science
Field: Social Sciences
Domain: Social Sciences
Open Access status: hybrid
OpenAlex ID: W7166325921
