A new model created by researchers at the University of Jyväskylä, Finland, allows computers to read and comprehend human emotions.
“Humans naturally interpret and react to each other’s emotions, a capability that machines fundamentally lack,” said Jussi Jokinen, Associate Professor of Cognitive Science. “This discrepancy can frustrate interactions with computers, especially if the machine remains oblivious to the user’s emotional state. Our model can be integrated into AI systems, granting them the ability to psychologically understand emotions and thus better relate to their users.”
Thanks to this model, computers may be able to predict human behavior in the future, including when a user might get irritated or anxious. In certain cases, the computer might provide further guidance or reroute the conversation. The new model can identify a variety of feelings, such as joy, boredom, annoyance, anger, despair, and worry.
“Consider a computer error during a critical task. The user’s cognition assesses this event as being counterproductive,” Jokinen explained. “An inexperienced user might react with anxiety and fear due to uncertainty on how to resolve the error, whereas an experienced user might feel irritation, annoyed at having to waste time resolving the issue. Our model predicts the user’s emotional response by simulating this cognitive evaluation process.”
According to the researchers, the model may anticipate user suffering and attempt to lessen unpleasant feelings. They emphasized how careful handling of emotional dynamics might enhance user experience in various contexts, including social media platforms and business settings.
Researchers underlined the dangers of an AI system not predicting users’ emotional reactions to interactive events, which could lead to a misalignment between humans and AI. Thanks to computational cognitive models such as the one created in this study, artificial agents can be made to grasp emotions clearly. This understanding extends beyond the mere ability to forecast feelings from behaviour or physiological cues. To enable seamless interactions with the user, it comprises internally modelling multiple “what if” scenarios and offering explanations for the origins of anticipated feelings.