Gastroenterology ·
ATTEST Framework Proposes Risk-Tiered AI Disclosure for Research
A special article in *Hepatology* proposes ATTEST, a framework for reporting AI use according to its potential effect on evidence generation. It assigns AI functions to four risk tiers and proposes a submission checklist and reporting standards for authors, reviewers, and editors.
- ATTEST classifies AI use across eight functional categories.
- Each category is assigned one of four risk tiers.
- Authors would complete a checklist at submission.
- The framework has not been validated in the supplied text.
Artificial intelligence is used in research tasks ranging from grammar editing to data extraction, pathology classification, and causal inference. The authors argue that generic disclosures do not distinguish these uses or show how AI may have influenced evidence generation. In this special article, they propose ATTEST (AI Transparency Tiered by Evidentiary Stakes), a reporting framework for research workflows.
ATTEST groups AI assistance into eight broad functional categories and assigns each a risk tier: Low, Medium, High, or Very High. The framework is intended to be implemented through a declaration checklist completed by authors at submission. It also proposes minimum reporting standards for each function, cross-cutting requirements, and obligations for reviewers and editors.
The proposal aims to help readers and journal teams assess AI’s role in scientific findings. It is a framework article, not an evaluation showing that the checklist improves research integrity or reproducibility. The supplied text does not provide validation data or report adoption by journals or professional organizations; the proposed standards therefore represent recommendations from the authors rather than established practice.
Is this summary clinically accurate?
Help fellow clinicians: your rating sends inaccurate summaries straight to our editors.
Sign in to rate this summary →Source
Hepatology: ATTEST 2027: A Framework for AI Transparency Tiered by Evidentiary Stakes ↗This is an automated AI-condensed summary that has not yet been reviewed by an editor. Always consult the full item at the original source.