# Don't Lie to Me: Avoiding Malicious Explanations with STEALTH

arXiv: https://arxiv.org/abs/2301.10407 · HTML: https://mlsanity.com/paper/2301.10407/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 15 of its 34 references that are in the map.
- First posted on arXiv 25 Jan 2023.

## Builds on (most cited first, 15 of 15)

- ["Why Should I Trust You?": Explaining the Predictions of Any Classifier](https://mlsanity.com/paper/1602.04938/) (2016)
- [Fooling LIME and SHAP: Adversarial Attacks on Post hoc Explanation Methods](https://mlsanity.com/paper/1911.02508/) (2019)
- [Explaining by Removing: A Unified Framework for Model Explanation](https://mlsanity.com/paper/2011.14878/) (2020)
- [Counterfactual Explanations Can Be Manipulated](https://mlsanity.com/paper/2106.02666/) (2021)
- [Bias in Machine Learning Software: Why? How? What to do?](https://mlsanity.com/paper/2105.12195/) (2021)
- [General Pitfalls of Model-Agnostic Interpretation Methods for Machine Learning Models](https://mlsanity.com/paper/2007.04131/) (2020)
- [Can I Trust My Fairness Metric? Assessing Fairness with Unlabeled Data and Bayesian Inference](https://mlsanity.com/paper/2010.09851/) (2020)
- [FairMask: Better Fairness via Model-based Rebalancing of Protected Attributes](https://mlsanity.com/paper/2110.01109/) (2021)
- [Deceptive AI Explanations: Creation and Detection](https://mlsanity.com/paper/2001.07641/) (2020)
- [Improving LIME Robustness with Smarter Locality Sampling](https://mlsanity.com/paper/2006.12302/) (2020)
- [Select Wisely and Explain: Active Learning and Probabilistic Local Post-hoc Explainability](https://mlsanity.com/paper/2108.06907/) (2021)
- [Unfooling Perturbation-Based Post Hoc Explainers](https://mlsanity.com/paper/2205.14772/) (2022)
- ["Sampling"' as a Baseline Optimizer for Search-based Software Engineering](https://mlsanity.com/paper/1608.07617/) (2016)
- [Better sampling in explanation methods can prevent dieselgate-like deception](https://mlsanity.com/paper/2101.11702/) (2021)
- [An Exploratory Study on Machine Learning Model Stores](https://mlsanity.com/paper/1905.10677/) (2019)

## Built on by (most cited first, 2 of 2)

- [Fairness Testing through Extreme Value Theory](https://mlsanity.com/paper/2501.11597/) (2025)
- [Software Fairness Dilemma: Is Bias Mitigation a Zero-Sum Game?](https://mlsanity.com/paper/2508.03323/) (2025)

## Abstract (arXiv)

STEALTH is a method for using some AI-generated model, without suffering from malicious attacks (i.e. lying) or associated unfairness issues. After recursively bi-clustering the data, STEALTH system asks the AI model a limited number of queries about class labels. STEALTH asks so few queries (1 per data cluster) that malicious algorithms (a) cannot detect its operation, nor (b) know when to lie.