Setting $\varepsilon$ is not the Issue in Differential Privacy

TL;DR

This paper contends that the challenge of setting the privacy budget in differential privacy is often overstated, and that interpreting privacy risks is a broader issue not unique to differential privacy, emphasizing the importance of rigorous risk assessment methods.

Contribution

It clarifies that the difficulty in setting privacy budgets is due to contextual risk estimation challenges, not the differential privacy framework itself, and advocates for sound risk assessment methods within the framework.

Findings

Interpreting privacy budgets is a broader risk estimation issue.

Differential privacy's definition is not the core problem.

Sound privacy risk assessment methods should align with differential privacy.

Abstract

This position paper argues that setting the privacy budget in differential privacy should not be viewed as an important limitation of differential privacy compared to alternative methods for privacy-preserving machine learning. The so-called problem of interpreting the privacy budget is often presented as a major hindrance to the wider adoption of differential privacy in real-world deployments and is sometimes used to promote alternative mitigation techniques for data protection. We believe this misleads decision-makers into choosing unsafe methods. We argue that the difficulty in interpreting privacy budgets does not stem from the definition of differential privacy itself, but from the intrinsic difficulty of estimating privacy risks in context, a challenge that any rigorous method for privacy risk assessment face. Moreover, we claim that any sound method for estimating privacy risks should, given the current state of research, be expressible within the differential privacy framework or justify why it cannot.