The Ethics of Emotional Artificial Intelligence: A Mixed Method Analysis

Area of Science:

• Psychology and Computer Science
• Human-Computer Interaction
• AI Ethics

Background:

• Emotions significantly influence human interactions, decisions, and actions.
• Artificial intelligence (AI) is being developed to interpret human emotions and predict behavior.
• Accurately measuring and evaluating emotions, even with AI, remains a significant challenge.

Purpose of the Study:

• To examine emotional variables and perceptions of emotional AI among college students.
• To investigate the ethical implications of using AI in emotion recognition.
• To analyze potential biases and discrimination arising from affective computing.

Main Methods:

• Quantitative and qualitative analysis of survey data from college students at an international university in Japan.
• Surveys included multiple-choice and open-ended essay questions on emotional AI.
• Participants represented diverse Asian nationalities, including Japanese, Indonesian, Korean, Chinese, Thai, and Vietnamese.

Main Results:

• Affective computing and data correlation for individual classification can yield ethically questionable results, potentially increasing bias and discrimination.
• Quantitative analysis highlighted risks associated with AI-driven emotion recognition.
• Qualitative analysis of student essays revealed an optimistic outlook on emotional AI applications.

Conclusions:

• The study underscores the ethical risks of AI in emotion recognition, particularly concerning bias and discrimination.
• Despite potential pitfalls, students expressed optimism regarding emotional AI's future applications.
• Increased awareness of the ethical challenges in emotional AI is essential for responsible development and deployment.