- Who it helps
- Researchers and educators evaluating AI-assisted learning and verification.
- Availability
- The public paper was submitted in August and revised September 5. This entry follows version 5; it is a research framework, with no ready-to-run product verified.
Separate immediate help from later learning
The proposed protocol compares delayed judgments after different practice conditions. A separate measure compares immediate performance with and without assistance. An immediate drop cannot by itself establish lost skill or dependence.
Leave uncertain outcomes unresolved
The revised paper explains why an uncertain delayed comparison does not establish a clean transfer pattern. Useful evaluation requires held-out claims and explicit comparison conditions.
What you can do
- Read version 5’s definitions and uncertainty treatment before adapting the protocol to a learning study. Open reference ↗
Where the evidence stops
The revised abstract and history were read. No study was reproduced, and lasting learning benefits are not established here.
Sources read. Sources checked October 21, 2026. No product workflow or benchmark was executed.
Sources & further reading
Paper · paper
Epistemic Transfer in AI-Assisted Verification, version 5 ↗
The revised abstract distinguishes later learning from the immediate effect of removing a tool.
Christoph Trattner · page checked 2026-10-21 · Published 2026-09-05
Coverage note: Independent hands-on testing and cross-platform discussion are incomplete. No missing links, creator identities or engagement counts are inferred.
The record over time (2)
Dated source checks and published explanations. An observation date can differ from the original release date.
- · Source recheck
Epistemic transfer asks whether practice with an AI fact-checking tool changes how well people judge unfamiliar claims later.
- · Added to Current
Epistemic transfer asks whether practice with an AI fact-checking tool changes how well people judge unfamiliar claims later.