Australia develops AI that can understand human emotions

The researchers said they trained the AI ​​by presenting sets of relevant facial expressions, allowing the machine to interpret emotions in a richer, more human-like context.

Illustration: aicompetence.org

Australian researchers have published a method to help artificial intelligence (AI) systems better perceive emotions, laying the foundation for more effective and empathetic human-machine interactions.

In a statement posted on the Science Media Exchange website on June 4, Edith Cowan University (ECU) in Western Australia said that although it has not yet been applied to physical robots, this research could soon influence emotional intelligence systems in the fields of mental health support, customer service andeducation worldwide.

The ECU researchers did not use the usual technique of analyzing single facial images. They trained the AI ​​by presenting sets of related facial expressions, allowing the machine to interpret emotions in a richer, more human-like context.

“Just like we don’t judge a person’s emotions from just one look, our method uses multiple expressions to make more informed predictions,” said lead author of the study, ECU PhD student Sharjeel Tahir.

This method increases reliability, even when faces appear at different angles or under different lighting, according to Tahir.

Exposing AI to diverse visual cues not only increases accuracy but also maintains the technology's effectiveness, said study co-author Dr Nima Mirnateghi of ECU.

The team is currently focused on developing artificial empathy, allowing AI agents to respond appropriately to human emotions and making AI decision-making more transparent and understandable.

The research team leader, ECU senior lecturer and AI expert Syed Afaq Shah said the work is laying the foundation for machines that can truly understand and respond to human emotions.

Back to topic

VNA

Source: https://tuoitre.vn/uc-phat-trien-ai-co-the-hieu-cam-cuc-con-nguoi-20250606084619139.htm