Inducing Uncertainty on Open-Weight Models for Test-Time Privacy in Image Recognition
Abstract
A key concern for AI safety remains understudied in the machine learning (ML) literature: how can we ensure users of ML models do not leverage predictions on incorrect personal data to harm others? This is particularly pertinent given the rise of open-weight models, where simply masking model outputs does not suffice to prevent adversaries from recovering harmful predictions. To address this threat, which we call *test-time privacy*, we induce maximal uncertainty on protected instances while preserving accuracy on all other instances. Our proposed algorithm uses a Pareto optimal objective that explicitly balances test-time privacy against utility. We also provide a certifiable approximation algorithm which achieves guarantees without convexity assumptions. We then prove a tight bound that characterizes the privacy-utility tradeoff that our algorithms incur. Empirically, our method obtains at least stronger uncertainty than pretraining with marginal drops in accuracy on various image recognition benchmarks. Altogether, this framework provides a tool to guarantee additional protection to end users.
Cite
@article{arxiv.2509.11625,
title = {Inducing Uncertainty on Open-Weight Models for Test-Time Privacy in Image Recognition},
author = {Muhammad H. Ashiq and Peter Triantafillou and Hung Yun Tseng and Grigoris G. Chrysos},
journal= {arXiv preprint arXiv:2509.11625},
year = {2025}
}