Affective Computing and Emotional Data: Challenges and Implications in Privacy Regulations, The AI Act, and Ethics in Large Language Models

Emotional data is increasingly collected without explicit consent: The paper distinguishes between emotional data gathered with informed consent in research settings versus data passively collected through everyday digital interactions (like monitoring facial expressions or speech patterns). This implicit collection raises significant privacy concerns that current regulations may not adequately address.

Emotion recognition systems have documented biases across demographic groups: The research highlights that AI systems trained to recognize emotions show varying accuracy and reliability depending on cultural background and demographic characteristics, meaning these systems may misinterpret or discriminate against certain populations—a critical issue as these technologies expand into healthcare, education, and customer service.

Emotional data should be treated as sensitive personal information requiring strong legal protections: The paper argues that emotional data qualifies as sensitive personal data under GDPR and should receive comparable safeguards to health or biometric information, yet current regulatory frameworks may not provide adequate protection for this uniquely intimate category of personal information.

LLMs like ChatGPT and Claude can now recognize and respond to human emotions, creating new ethical risks: As large language models become more sophisticated at detecting emotional states through text, they create novel scenarios where AI systems can influence human behavior based on emotional vulnerabilities—raising questions about manipulation and individual autonomy that existing AI ethics frameworks haven't fully addressed.

Cultural differences in emotional expression complicate global AI deployment: The paper emphasizes that emotions are expressed differently across cultures, yet emotion recognition systems are often trained on limited datasets, risking ineffective or inappropriate responses when deployed internationally.