SkillBot: Identifying Risky Content for Children in Alexa Skills

TL;DR

This paper introduces SkillBot, an NLP system to detect risky child-directed Alexa skills, revealing vulnerabilities like inappropriate content, privacy issues, and confounding utterances that pose risks to children despite stricter policies.

Contribution

We develop SkillBot to automatically analyze Alexa skills, identify risky content, and uncover confounding utterances, providing new insights into vulnerabilities in child-focused voice apps.

Findings

Identified 28 risky child-directed Alexa skills.

Collected 31,966 app behaviors from 3,434 skills.

Found 4,487 confounding utterances, 27% of which favor non-child apps.

Abstract

Many households include children who use voice personal assistants (VPA) such as Amazon Alexa. Children benefit from the rich functionalities of VPAs and third-party apps but are also exposed to new risks in the VPA ecosystem. In this paper, we first investigate "risky" child-directed voice apps that contain inappropriate content or ask for personal information through voice interactions. We build SkillBot - a natural language processing (NLP)-based system to automatically interact with VPA apps and analyze the resulting conversations. We find 28 risky child-directed apps and maintain a growing dataset of 31,966 non-overlapping app behaviors collected from 3,434 Alexa apps. Our findings suggest that although child-directed VPA apps are subject to stricter policy requirements and more intensive vetting, children remain vulnerable to inappropriate content and privacy violations. We then conduct a user study showing that parents are concerned about the identified risky apps. Many parents do not believe that these apps are available and designed for families/kids, although these apps are actually published in Amazon's "Kids" product category. We also find that parents often neglect basic precautions such as enabling parental controls on Alexa devices. Finally, we identify a novel risk in the VPA ecosystem: confounding utterances, or voice commands shared by multiple apps that may cause a user to interact with a different app than intended. We identify 4,487 confounding utterances, including 581 shared by child-directed and non-child-directed apps. We find that 27% of these confounding utterances prioritize invoking a non-child-directed app over a child-directed app. This indicates that children are at real risk of accidentally invoking non-child-directed apps due to confounding utterances.