# The Audio Auditor: User-Level Membership Inference in Internet of Things Voice Services

## TL;DR

This paper introduces an audio auditor that can determine if a user's voice data was used in training IoT voice recognition models, achieving over 80% accuracy on real-world Siri data, thus highlighting privacy risks.

## Contribution

The paper presents a novel user-level membership inference method for IoT voice services that generalizes across different ASR architectures and real-world applications.

## Key findings

- Auditor achieves over 80% accuracy on Siri data.
- Auditor generalizes well across different ASR models.
- Method informs privacy protection strategies for IoT devices.

## Abstract

With the rapid development of deep learning techniques, the popularity of voice services implemented on various Internet of Things (IoT) devices is ever increasing. In this paper, we examine user-level membership inference in the problem space of voice services, by designing an audio auditor to verify whether a specific user had unwillingly contributed audio used to train an automatic speech recognition (ASR) model under strict black-box access. With user representation of the input audio data and their corresponding translated text, our trained auditor is effective in user-level audit. We also observe that the auditor trained on specific data can be generalized well regardless of the ASR model architecture. We validate the auditor on ASR models trained with LSTM, RNNs, and GRU algorithms on two state-of-the-art pipelines, the hybrid ASR system and the end-to-end ASR system. Finally, we conduct a real-world trial of our auditor on iPhone Siri, achieving an overall accuracy exceeding 80\%. We hope the methodology developed in this paper and findings can inform privacy advocates to overhaul IoT privacy.