Artificial intelligence and social intelligence: preliminary comparison study between AI models and psychologists

Background: Social intelligence (SI) is crucial for success in counseling and psychotherapy for both human psychologists and AI systems assisting them. This study aimed to identify the SI of artificial intelligence, represented by large language models (ChatGPT-4, Google Bard, Bing), compared to psychologists. Methods: A stratified random sample of 180 counseling psychology students (bachelor’s and doctoral levels) from King Khalid University participated; AI models included ChatGPT-4, Google Bard, and Bing. Both psychologists and AI models completed the Social Intelligence Scale. Results: Significant differences in SI were found between psychologists and AI models ChatGPT-4 and Bing. ChatGPT-4 exceeded 100% of all psychologists, and Bing outperformed 50% of PhD holders and 90% of bachelor’s holders. Differences between Google Bard and bachelor students were not significant; however, 90% of PhD holders outperformed Google Bard. Conclusion: The study explored applying human measures to AI entities and found rapid development in AI’s understanding of emotions and social behavior related to SI. AI is poised to substantially assist psychotherapists, and clinicians should anticipate further AI development and benefits for counseling and psychotherapy. Further studies using humanistic and non-humanistic criteria with large language models are needed.

Nabil Saleh Sufyan, Fahmi H. Fadhel, Saleh Safeer Alkhathami, Jubran Y. A. Mukhadi