Legal Obligation and Ethical Best Practice: Towards Meaningful Verbal Consent for Voice Assistants

Voice User Interface (VUI) DesignPrivacy by Design & User ControlResearch Ethics & Open ScienceAI/ML Researchers & EngineersLawyers & Legal ResearchersPrivacy Policy MakersHCI Researchers

Title of the Paper

Legal Obligation and Ethical Best Practice: Towards Meaningful Verbal Consent for Voice Assistants

Paper Information

  • Subject Area: Data Privacy, Voice Assistants, User Experience Design
  • Keywords: Voice Assistants, Consent, Verbal Consent, Informed Consent, GDPR, Alexa, Conversational User Interface, Permissions

Research Background and Issues

  • Challenges and Issues:

    1. The data collection and user consent mechanisms of voice assistants (VAs) often rely on mobile app models, but the usability of this model faces challenges in voice assistant scenarios.
    2. Platforms like Alexa currently provide voice-based "Voice-Forward Consent" (VFC), but this approach may conflict with existing principles of informed consent (voluntariness, specificity, informedness, revocability, and non-burdensomeness).
    3. There is ongoing debate regarding the ethical soundness and legal compliance of current voice-based consent mechanisms.
  • Significance of the Research: As voice assistants become increasingly integrated into daily household life, their interaction methods directly impact user privacy. Ensuring that permission mechanisms for voice assistants comply with legal requirements while reflecting ethical and user-friendly design is of paramount importance.

  • Motivation and Related Work:

    • This paper aims to explore how to achieve "meaningful verbal consent" in voice assistants that satisfies both legal requirements and ethical best practices.
    • It evaluates existing verbal consent technologies by combining the legal requirements of the General Data Protection Regulation (GDPR) with research in the field of human-computer interaction.

Proposed Solution

  • Proposed Solution: Using the Delphi method, the study conducted multiple rounds of surveys with experts from academia, industry, and policy-making fields to evaluate the key requirements for verbal consent mechanisms. These requirements were derived from literature, regulations, and ethical guidelines.

  • Innovative Aspects of the Solution:

    • Identified seven highly relevant and critical implementation requirements for verbal consent in voice assistants.
    • Proposed six long-term recommendations to advance "meaningful verbal consent," covering platform architecture, privacy rules, and user experience design considerations.
    • Innovatively integrated user habits and implicit social ethics into voice consent mechanisms to reduce user "consent fatigue."
  • Implementation Steps and Key Techniques:

    1. Requirement Extraction: Derived 41 requirements for verbal consent in voice assistants from existing literature, regulations (e.g., GDPR), and ethical guidelines.
    2. Delphi Survey: Conducted multiple rounds of questionnaires where experts evaluated the relevance, feasibility, and usability of the requirements and discussed existing disagreements.
    3. Analysis and Recommendations:
      • Prioritized different requirements and proposed specific improvements and long-term directions.
      • Designed reasonable suggestions for integrating GDPR compliance and user needs into the permission management process of voice assistants.

Research Findings

  • Specific Findings:

    • Identified the following seven key requirements:
      1. Verbal consent should explain how to revoke permissions.
      2. Platforms should ensure users can distinguish and trust voices originating from the platform rather than third-party skills.
      3. The consent process should require explicit and affirmative statements from users.
      4. Users should be able to revoke skill access permissions using voice commands.
      5. Clear prompts should indicate whether data is used for platform or internet tracking.
      6. Skill developers must publish privacy policies.
      7. Platforms should periodically verify the validity of third-party privacy policies.
    • Proposed six recommendations for improving existing VFC systems, including separating different legal bases for data use, reducing the frequency of user consent decisions, and enhancing overall transparency.
  • Experimental and Evaluation Results:

    • Pain Points: Current detailed explanations of user privacy are often "too lengthy or difficult to understand," and the direct adaptation to GDPR incurs high costs.
    • Advantages: Through expert feedback and theoretical advancements, the study provides practical and feasible solutions to reduce user burden and enhance user control.
  • Limitations and Future Directions:

    1. This study is based on the GDPR framework in the EU, which may limit its applicability to other legal frameworks.
    2. The improvement proposals derived from expert surveys require further validation of actual usability through user experiments.

Quick Actions

Share

Share this page

ios_share

https://hci.top/en/papers/chi/96279/2023

AdRecommended

Learn AI Coding at CodeNow

open_in_newOpen DOI Link
DOI: https://doi.org/10.1145/3544548.3580967
At a Glance

Paper Snapshot

fact_check
dataset
Source
CHI
calendar_month
Year
2023
emoji_events
Award
No award tagged
group
Authors
3 authors
sell
Subtopics
Voice User Interface (VUI) Design, Privacy by Design & User Control, Research Ethics & Open Science
work
Professions
AI/ML Researchers & Engineers, Lawyers & Legal Researchers, Privacy Policy Makers, HCI Researchers
article
Content Status
Full text indexed
hub
Related Papers
1 related papers