Independent veterinary AI evaluation
Big claims. Show the evidence.
Published Paper Checks and Claim Checks separate what veterinary AI evidence supports from what remains unverified.
Latest evidence
Our first Paper Check audits the boundary between scarce public disclosure and the stronger claim that validation never happened.
A systematic audit found very little validation, safety or data-provenance information in public vendor materials. It measured disclosure—not product performance—and cannot show that validation never happened.
Latest Claim Check
We preserve the commercial claim, the study result and the editorial assessment as distinct layers.
Partially supported — equivalence not demonstrated
The reported accuracy point estimates were close: 90.2% for SignalRAY and 91.0% for the best participating radiologist. But the study tested for a difference; it did not use a prespecified equivalence or non-inferiority framework capable of demonstrating equivalent performance.
Research Radar
Research Radar watches veterinary-AI papers, validation studies, vendors, products, commercial claims, meaningful claim changes and evidence gaps.
Research Radar findings are not automatically published. Every public VetAI Trust assessment passes human editorial review.
How Research Radar worksMethod
A test is commissioned only where an evidence gap matters and an independent protocol could answer it honestly.
Preserve the original wording and define what it promises.
Map studies, sources, limitations and what has actually been reviewed.
Design an independent benchmark where existing evidence leaves a useful question open.
Publish what is supported, qualified and still unresolved.
Tests, trust and governance
VetAI Trust does not sell rankings. Funding or vendor participation must not control editorial conclusions.
Free benchmark submissions are open, but no result or ranking is implied before a disclosed protocol and publishable finding exist.
Governance rules separate financial support, vendor participation and editorial conclusions.
Material changes should remain visible instead of silently rewriting the record.
