Scholarlink

<p>Public perception of Artificial Intelligence (AI) risks is essential in shaping the governance of AI. This study examines societal concerns regarding AI emotion recognition systems, focusing on 28 feedback letters submitted to the European Union in response to a draft version of the AI Act. We conduct a thematic analysis of the arguments in these letters to uncover the underlying concerns that motivated stakeholder participation in the public consultation process. This analysis provides insight into the perceived risks, values, and societal expectations related to the regulation of emotion recognition technologies. The findings reveal that most concerns cluster around three values: human agency and oversight; technical robustness and safety; and diversity, non-discrimination and fairness. The risk of violation of these values align with the rationale in the AI Act underlying the restrictions or bans on emotion recognition systems. Our research sheds light on how public opinion influenced the AI Act and provides recommendations for future legislative and technological development in the field of AI emotion recognition.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Public Concerns about Emotion Recognition Systems: A Thematic Analysis

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月10日

查看全文
  • Banu Buruk,
  • Anne-Marie Brouwer,
  • Giulio Mecacci,
  • Deniz Iren,
  • Lucas Noldus,
  • Marc Steen,
  • Teun de Waal

摘要

Public perception of Artificial Intelligence (AI) risks is essential in shaping the governance of AI. This study examines societal concerns regarding AI emotion recognition systems, focusing on 28 feedback letters submitted to the European Union in response to a draft version of the AI Act. We conduct a thematic analysis of the arguments in these letters to uncover the underlying concerns that motivated stakeholder participation in the public consultation process. This analysis provides insight into the perceived risks, values, and societal expectations related to the regulation of emotion recognition technologies. The findings reveal that most concerns cluster around three values: human agency and oversight; technical robustness and safety; and diversity, non-discrimination and fairness. The risk of violation of these values align with the rationale in the AI Act underlying the restrictions or bans on emotion recognition systems. Our research sheds light on how public opinion influenced the AI Act and provides recommendations for future legislative and technological development in the field of AI emotion recognition.