aceCan AI Predict Publication? Multimodal Large Language Models and the Structural Determinants of Surgical Scholarship
BackgroundWhether artificial intelligence can identify publishable scientific work is untested. We evaluated whether a multimodal large language model (MLLM) could predict, from poster content alone, which abstracts at the American Association for the Surgery of Trauma (AAST) Annual Meetings reached publication, and characterized the investigator, institutional, and domain level determinants situating model performance.MethodsWe retrospectively analyzed 260 abstracts from the 2021-2022 AAST Annual Meetings. Bibl…
GPT-4.1 scored poster images alone and predicted which AAST abstracts reached publication with 58.5% overall accuracy, rising to 74.2% in Violence, Societal, and Behavioral and 62.2% in Hemorrhage, Resuscitation, and Vascular Control.
Model performance fell to chance in Critical Care and Outcomes and Systems, Technology, and Process Optimization, where publication depended on institutional factors absent from the poster such as multicenter scaffolding, mentorship, and senior author fluency.
Findings limited to 260 AAST abstracts from 2021-2022, with investigator demographics inferred from public data and model performance dependent on domain and on structural factors not visible on posters.
Evidence
- Peer-reviewedThe American Surgeon™2026-09-14
How should this claim be treated?
Truvace Impact Record TRV-2026-1114, v1: “Can AI Predict Publication? Multimodal Large Language Models and the Structural Determinants of Surgical Scholarship.” Truvace, 2026-09-16. /record/TRV-2026-1114 (accessed at citation time). sha256 4b63f26ccef6e9c6…
Calibration history
Every change to this record since certification, in the open. None yet — the reading has held since it entered the record.
Certified into the record
How to verify without trusting this page
Fetch the canonical text of any version from /api/record/TRV-2026-1114 and hash it yourself — for example shasum -a 256 on the saved canonical field. The result must equal content_hash, and each version’s text ends with prev:followed by the prior version’s hash (version 1 chains to 64 zeros). If a single character of any version had been altered since certification, the chain would not reproduce.