Finnish study: AI could aid education and healthcare by reading behaviour

Artificial intelligence could soon play a role in education, healthcare and other fields where reading social signals is critical, according to new research from the University of Turku.

A team at the Turku PET Centre found that ChatGPT, the large language model developed by OpenAI, was able to evaluate subtle social features in images and videos with a consistency that sometimes surpassed human participants.

The study asked the model to assess 138 different social cues, such as cooperation, emotional expression, posture and facial gestures. Its answers were then compared with those of more than 2 000 people.

“Since ChatGPT's assessment of social features were on average more consistent than those of an individual participant, its evaluations could be trusted even more than those made by a single person,” said postdoctoral researcher Severi Santavirta from the University of Turku.

“However, the evaluations of several people together are still more accurate than those of artificial intelligence.”

Efficiency and future applications in education and healthcare

According to Santavirta, the model’s output closely mirrored human data, with results proving strikingly similar when the team mapped the brain networks of social perception based on either ChatGPT’s evaluations or those of people.

One of the most notable advantages of this approach is efficiency. 

“Collecting human evaluations required the efforts of more than 2 000 participants and a total of more than 10 000 work hours, while ChatGPT produced the same evaluations in just a few hours.”

Looking ahead, Santavirta suggested AI could take on a supportive monitoring role in classrooms and clinics, allowing humans to focus on confirming the most important observations. “The AI does not get tired like a human, but can monitor situations around the clock.”