Tubes Among Us: Analog Attack on Automatic Speaker Identification

TL;DR

This paper reveals that humans can easily create analog adversarial examples for speaker identification by speaking through a tube, challenging assumptions about defenses against machine learning-based audio attacks.

Contribution

It demonstrates that humans can produce effective analog adversarial examples for speaker identification, undermining defenses that assume human inability to generate such attacks.

Findings

Humans can impersonate speakers by speaking through a tube.

Analog attacks can bypass machine learning defenses in speaker ID.

Implications for security-critical acoustic biometric systems.

Abstract

Recent years have seen a surge in the popularity of acoustics-enabled personal devices powered by machine learning. Yet, machine learning has proven to be vulnerable to adversarial examples. A large number of modern systems protect themselves against such attacks by targeting artificiality, i.e., they deploy mechanisms to detect the lack of human involvement in generating the adversarial examples. However, these defenses implicitly assume that humans are incapable of producing meaningful and targeted adversarial examples. In this paper, we show that this base assumption is wrong. In particular, we demonstrate that for tasks like speaker identification, a human is capable of producing analog adversarial examples directly with little cost and supervision: by simply speaking through a tube, an adversary reliably impersonates other speakers in eyes of ML models for speaker identification. Our findings extend to a range of other acoustic-biometric tasks such as liveness detection, bringing into question their use in security-critical settings in real life, such as phone banking.