ChatGPT Virtual Assistant for Breast Reconstruction: Assessing Preferences for a Traditional Chatbot versus a Human AI VideoBot

BACKGROUND: The recent advancements in artificial intelligence (AI) has shifted the landscape of telehealth in which AI chatbots, such as ChatGPT, have demonstrated numerous potential applications in medicine (1). While text-based systems are efficient and increase the availability of health information, they are not a personal approach to patient interaction and lack human-like qualities that increase user effectiveness, usability, and trust (2). In the field of plastic and reconstructive surgery (PRS), ChatGPT has shown to be capable of providing basic, general patient education on topics such as rhinoplasty and breast augmentation (3, 4). However, preferences for a traditional chatbot versus a human AI VideoBot in the context of PRS have yet to be assessed. METHODS: Synthesia, an AI video generation platform, was utilized to develop a human AI avatar that was integrated with ChatGPT 3.5 to create the VideoBot. The VideoBot was then integrated into Tolstoy, a platform for creating interactive videos, to develop an interactive VideoBot whereby the human AI avatar was able to answer four of the most commonly asked questions related to breast reconstruction. We utilized Zapier, a workflow automation software, to develop a ChatGPT 3.5-integrated text-based chatbot. An anonymous 20 item survey adapted from Corritore et al.‘s 2005 validated Measurement of Online Trust (5) was distributed online via Amazon Mechanical Turk to female participants. Participant demographic data was collected, and participants were asked to rank their preferences between the VideoBot and text-based chatbot on a 1-7 Likert scale (1=Strongly Disagree, 7= Strongly Agree). RESULTS: A total of 396 responses were gathered. Participants were 18-64 years old, 97% had received some form of plastic or reconstructive surgery, and 95% of those patients underwent a breast-related medical procedure. Perceptions of truthfulness, believability, content expertise, ease of use, and safety were similar between the VideoBot and chatbot. Most participants preferred the VideoBot over the traditional chatbot (63.5% versus 28.1%) as they found it more captivating than the text-based chatbot. 77% of participants would have preferred to see someone that they identified with as the human AI avatar in terms of gender and race. CONCLUSION: There is no difference in user effectiveness, usability, and trust between the VideoBot and text-based chatbot. However, the human-like quality of the VideoBot allows for a more interactive experience than the traditional chatbot. The race and gender effect should be further explored in providing a more personable telehealth experience for patients. REFERENCES: 1. Thirunavukarasu AJ, Ting DSJ, Elangovan K, Gutierrez L, Tan TF, Ting DSW. Large language models in medicine. Nat Med. Aug 2023;29(8):1930-1940. doi:10.1038/s41591-023-02448-8 2. F. Dsouza, R. Shaharao, Y. Thakur, P. Agwan, G. Sakarkar and P. Gupta, “Advancement in Communication using Natural Language based VideoBot System,” 2022 IEEE Bombay Section Signature Conference (IBSSC), Mumbai, India, 2022, pp. 1-5, doi: 10.1109/IBSSC56953.2022.10037380. 3. Xie Y, Seth I, Hunter-Smith DJ, Rozen WM, Ross R, Lee M. Aesthetic Surgery Advice and Counseling from Artificial Intelligence: A Rhinoplasty Consultation with ChatGPT. Aesthetic Plast Surg. Oct 2023;47(5):1985-1993. doi:10.1007/s00266-023-03338-7 4. Seth I, Cox A, Xie Y, et al. Evaluating Chatbot Efficacy for Answering Frequently Asked Questions in Plastic Surgery: A ChatGPT Case Study Focused on Breast Augmentation. Aesthet Surg J. Sep 14 2023;43(10):1126-1135. doi:10.1093/asj/sjad140 5. Corritore CL, Marble RP, Wiedenbeck S, Kracher B, Chandran A. Measuring online trust of websites: Credibility, perceived ease of use, and risk.