Position: The Open Benchmark Paradox Must Be Resolved through Sovereign Medical Evaluation

Keonwoo Kim, Hyeseon Ko, Hyejeong Jo, Sewon Kim, Yera Choi, Jaedeok Lee, Heeyoung Kwak, Yunwook Sung, Haanju Yoo
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:171059-171071, 2026.

Abstract

As generative medical large language models become increasingly involved in clinical actions, public benchmarks are often treated as proxies of deployment-readiness. However, this reliance creates a false sense of security because public scores are often based on data the models have already seen. We call this the Open Benchmark Paradox: making evaluation data public for research progress also makes data contamination inevitable, ruining its value as a reliable safety signal. This paradox induces three structural failures: (1) hidden contamination, where it is impossible to prove evaluation independence; (2) outdated standards, where static datasets fail to track evolving medical guidelines; and (3) jurisdictional divergence, where global averaging ignores local legal and ethical standards. To validate these risks, we audited frontier models using recent medical exam data, which confirmed a high probability of data contamination. To resolve such integrity issues in medical evaluation, we propose Sovereign Medical Evaluation (SME). Instead of public leaderboards, SME establishes a national infrastructure where health authorities manage private, isolated evaluation pipelines. Within this secure system, evaluations are automatically updated using live medical data and legal changes, ensuring they remain current and strictly separated from model training. SME provides the essential transition to a controlled, auditable, and legally grounded safety gate for clinically deployed generative medical LLMs.

Cite this Paper


BibTeX
@InProceedings{pmlr-v306-kim26cd, title = {Position: The Open Benchmark Paradox Must Be Resolved through Sovereign Medical Evaluation}, author = {Kim, Keonwoo and Ko, Hyeseon and Jo, Hyejeong and Kim, Sewon and Choi, Yera and Lee, Jaedeok and Kwak, Heeyoung and Sung, Yunwook and Yoo, Haanju}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {171059--171071}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/kim26cd/kim26cd.pdf}, url = {https://proceedings.mlr.press/v306/kim26cd.html}, abstract = {As generative medical large language models become increasingly involved in clinical actions, public benchmarks are often treated as proxies of deployment-readiness. However, this reliance creates a false sense of security because public scores are often based on data the models have already seen. We call this the Open Benchmark Paradox: making evaluation data public for research progress also makes data contamination inevitable, ruining its value as a reliable safety signal. This paradox induces three structural failures: (1) hidden contamination, where it is impossible to prove evaluation independence; (2) outdated standards, where static datasets fail to track evolving medical guidelines; and (3) jurisdictional divergence, where global averaging ignores local legal and ethical standards. To validate these risks, we audited frontier models using recent medical exam data, which confirmed a high probability of data contamination. To resolve such integrity issues in medical evaluation, we propose Sovereign Medical Evaluation (SME). Instead of public leaderboards, SME establishes a national infrastructure where health authorities manage private, isolated evaluation pipelines. Within this secure system, evaluations are automatically updated using live medical data and legal changes, ensuring they remain current and strictly separated from model training. SME provides the essential transition to a controlled, auditable, and legally grounded safety gate for clinically deployed generative medical LLMs.} }
Endnote
%0 Conference Paper %T Position: The Open Benchmark Paradox Must Be Resolved through Sovereign Medical Evaluation %A Keonwoo Kim %A Hyeseon Ko %A Hyejeong Jo %A Sewon Kim %A Yera Choi %A Jaedeok Lee %A Heeyoung Kwak %A Yunwook Sung %A Haanju Yoo %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-kim26cd %I PMLR %P 171059--171071 %U https://proceedings.mlr.press/v306/kim26cd.html %V 306 %X As generative medical large language models become increasingly involved in clinical actions, public benchmarks are often treated as proxies of deployment-readiness. However, this reliance creates a false sense of security because public scores are often based on data the models have already seen. We call this the Open Benchmark Paradox: making evaluation data public for research progress also makes data contamination inevitable, ruining its value as a reliable safety signal. This paradox induces three structural failures: (1) hidden contamination, where it is impossible to prove evaluation independence; (2) outdated standards, where static datasets fail to track evolving medical guidelines; and (3) jurisdictional divergence, where global averaging ignores local legal and ethical standards. To validate these risks, we audited frontier models using recent medical exam data, which confirmed a high probability of data contamination. To resolve such integrity issues in medical evaluation, we propose Sovereign Medical Evaluation (SME). Instead of public leaderboards, SME establishes a national infrastructure where health authorities manage private, isolated evaluation pipelines. Within this secure system, evaluations are automatically updated using live medical data and legal changes, ensuring they remain current and strictly separated from model training. SME provides the essential transition to a controlled, auditable, and legally grounded safety gate for clinically deployed generative medical LLMs.
APA
Kim, K., Ko, H., Jo, H., Kim, S., Choi, Y., Lee, J., Kwak, H., Sung, Y. & Yoo, H.. (2026). Position: The Open Benchmark Paradox Must Be Resolved through Sovereign Medical Evaluation. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:171059-171071 Available from https://proceedings.mlr.press/v306/kim26cd.html.

Related Material