ChatGPT-4 Proves It Can Read Humans Better Than We Can

4/27/24

Editorial team at Bits with Brains

The integration of artificial intelligence (AI) and psychotherapy is no longer a futuristic concept—it's a present reality with profound implications.

A recent study published in “Frontiers in Psychology” has surfaced the exceptional performance of ChatGPT-4 in understanding and responding to human emotions and needs, a domain traditionally reserved for human psychologists.

The study's findings underscore the rapid advancements in AI capabilities, particularly in social intelligence. ChatGPT-4's score of 59 out of 64 on the Social Intelligence Scale is not just a number—it's a benchmark that surpasses the scores of all human psychologists who participated in the study. This is a significant development, considering that social intelligence is a cornerstone of effective psychotherapy.

The implications of such a development are far-reaching. For one, it sets a new standard for AI integration into mental health services. The performance of Bing, which outperformed over half of the psychologists, and Google Bard, which matched the performance of psychologists with bachelor's degrees, further highlights the potential of AI.

The study involved AI models and human participants responding to scenarios from the Social Intelligence Scale, with responses scored against predefined criteria. The results indicate that AI can match or even exceed human performance in social intelligence tasks. This sets a precedent for future AI applications in mental health services and highlights the need for careful consideration of AI's role in this area.

The ethical concerns surrounding AI in mental health are significant. There’s the question of whether AI can truly replace the empathetic and nuanced understanding that human psychologists offer. Additionally, the potential for AI to replace human jobs in psychology raises issues about the future of the profession and the quality of patient care.

Despite these concerns, the study's findings cannot be ignored. They suggest that AI has the potential to revolutionize mental health services by providing support that matches or exceeds human capabilities in understanding and responding to social cues. This could lead to more accessible and effective mental health care for patients, particularly in rural areas where there is a shortage of qualified professionals.

As organizations across various industries consider implementing or expanding their use of generative AI, it's crucial to recognize the transformative potential of such technology. For organizations looking to implement generative AI, the study serves as a reminder of the importance of approaching this technology with a balanced perspective. While AI offers innovative solutions and efficiencies, it also presents new ethical dilemmas and challenges that must be addressed thoughtfully.

Sources:

[1] https://arxiv.org/pdf/2204.09082.pdf

[2] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6532335/

[3] https://www.forbes.com/sites/bernardmarr/2023/07/06/ai-in-mental-health-opportunities-and-challenges-in-developing-intelligent-digital-therapies/?sh=58d754865e10

[4] https://www.youtube.com/watch?v=CnSBfHZv3b4

[5] https://www.infinitus.ai/blog/human-ai-collaboration/

[6] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10230127/

[7] https://www.reddit.com/r/Futurology/comments/1c99pee/chatgpt4_outperforms_human_psychologists_in_test/

[8] https://www.apa.org/news/podcasts/speaking-of-psychology/ethical-artificial-intelligence

[9] https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7334754/

[10] https://www.npr.org/sections/health-shots/2023/01/19/1147081115/therapy-by-chatbot-the-promise-and-challenges-in-using-ai-for-mental-health

[11] https://www.apa.org/monitor/2024/04/addressing-equity-ethics-artificial-intelligence