Russian Researchers Add Emotional Intelligence to Human-AI Dialogue
A new multimodal framework analyzes how emotions shape decisions in interactions between humans and machines.
Researchers at Moscow State University’s AI Center have developed a tool that helps analyze how emotions influence decision-making during interactions between humans and artificial intelligence systems.
Measuring Emotions
The system, called HL-EAI, is a software framework that integrates text, voice and video data to assess the emotional state of participants in a dialogue. Based on this analysis, the platform generates responses that take into account the conversational context.
The tool can be used to study communication across three formats: human-to-AI, human-to-human and even AI-to-AI interactions.
Bringing Emotions Together
Many existing AI systems can detect individual emotional signals from users. However, they treat emotions as isolated features and fail to capture how humans and machines “exchange feelings” in real conversations. The new architecture aims to bridge this gap.
A Platform for Experiments
So far, researchers are testing the system in game-based scenarios. Participants make decisions in various strategic and tactical situations, allowing scientists to analyze how emotional context influences trust, willingness to cooperate and overall behavioral patterns in interactions with algorithms.
