EAGER: SaTC: Towards Accounting for the Human in Emotion Recognition Technologies

点击翻译按钮获取中文摘要

Emotions are powerful, mediate human experiences with their surroundings, and impact decision-making and attention online and off. Privacy and emotion are related concepts; online and off, emotions are often deemed private. Sharing and signaling one's emotions to other humans can be beneficial, but involves privacy calculations and complex decision-making processes. Despite the deeply personal nature of human emotion, artificial intelligence (AI) algorithms are being built to recognize and infer emotions using data sources such as social media behavior, streaming service use, voice, facial expressions, biometrics, and body language in ways often unknown to users. Emotion recognition’s emerging market is expected to grow to $3.8 billion by 2025. This project posits that emotion recognition technologies can prioritize the privacy and preferences of the humans they impact. This project takes steps toward addressing users' privacy and other concerns by investigating people's attitudes towards emotion recognition and how users and technologists conceive of this technology. This project deepens intellectual conversations about emerging technologies' privacy and safety; interactions between ethics and AI, data science, and research; and social media design. Technologists, policy makers, and researchers in fields such as computing, economics, and medicine can use this project’s resulting framework to evaluate systems' potential implications. Furthermore, the project’s topic provides an important context to challenge students to think deeply about ethics, privacy, social responsibility, and technology in teaching and research — crucial for generating a responsible and thoughtful next generation of information and computing professionals.This work will focus on emotions and users' attitudes towards emotion recognition technologies to contribute to our knowledge about socially and ethically responsible use and treatment of data in algorithmic decision-making that impacts personal lives. Using a human-centered lens, this work will contribute a framework of guidelines within which emotion recognition technologies can be evaluated for compatibility with people's values, needs, and concerns. The research team will 1) conduct a series of critical analyses of artifacts, including patents, to interrogate the ways technologists and academics articulate emotion recognition technology's importance across application domains, and its use-cases, visions, and implications (e.g., ethical, privacy, personal, inter-personal, social); and 2) use interviews, vignettes, and focus groups to examine social media users' attitudes towards emotion recognition and its perceived impacts on their privacy and their lives, highlighting factors contributing to a trustworthy or untrustworthy cyberspace that may utilize emotion recognition.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.