Deep Adversarial Learning on Google Home devices

TL;DR

This paper evaluates the effectiveness of deep adversarial learning countermeasures against privacy attacks on Google Home devices, revealing their inadequacy and highlighting the need for re-engineering to enhance security.

Contribution

It provides an experimental assessment showing that current adversarial defenses are insufficient against machine learning-based privacy attacks on smart speakers.

Findings

Adversarial countermeasures fail to prevent traffic analysis attacks

Current defenses do not guarantee privacy for smart speaker users

Major re-engineering is required for effective protection

Abstract

Smart speakers and voice-based virtual assistants are core components for the success of the IoT paradigm. Unfortunately, they are vulnerable to various privacy threats exploiting machine learning to analyze the generated encrypted traffic. To cope with that, deep adversarial learning approaches can be used to build black-box countermeasures altering the network traffic (e.g., via packet padding) and its statistical information. This letter showcases the inadequacy of such countermeasures against machine learning attacks with a dedicated experimental campaign on a real network dataset. Results indicate the need for a major re-engineering to guarantee the suitable protection of commercially available smart speakers.