A Generative Pretrained Transformer (GPT)-Powered Chatbot as a Simulated Patient to Practice History Taking: Prospective, Mixed Methods Study.

BACKGROUND: Communication is a core competency of medical professionals and of utmost importance for patient safety. Although medical curricula emphasize communication training, traditional formats, such as real or simulated patient interactions, can present psychological stress and are limited in repetition. The recent emergence of large language models (LLMs), such as generative pretrained transformer (GPT), offers an opportunity to overcome these restrictions. OBJECTIVE: The aim of this study was to explore the feasibility of a GPT-driven chatbot to practice history taking, one of the core competencies of communication. METHODS: We developed an interactive chatbot interface using GPT-3.5 and a specific prompt including a chatbot-optimized illness script and a behavioral component. Following a mixed methods approach, we invited medical students to voluntarily practice history taking. To determine whether GPT provides suitable answers as a simulated patient, the conversations were recorded and analyzed using quantitative and qualitative approaches. We analyzed the extent to which the questions and answers aligned with the provided script, as well as the medical plausibility of the answers. Finally, the students filled out the Chatbot Usability Questionnaire (CUQ). RESULTS: A total of 28 students practiced with our chatbot (mean age 23.4, SD 2.9 years). We recorded a total of 826 question-answer pairs (QAPs), with a median of 27.5 QAPs per conversation and 94.7% (n=782) pertaining to history taking. When questions were explicitly covered by the script (n=502, 60.3%), the GPT-provided answers were mostly based on explicit script information (n=471, 94.4%). For questions not covered by the script (n=195, 23.4%), the GPT answers used 56.4% (n=110) fictitious information. Regarding plausibility, 842 (97.9%) of 860 QAPs were rated as plausible. Of the 14 (2.1%) implausible answers, GPT provided answers rated as socially desirable, leaving role identity, ignoring script information, illogical reasoning, and calculation error. Despite these results, the CUQ revealed an overall positive user experience (77/100 points). CONCLUSIONS: Our data showed that LLMs, such as GPT, can provide a simulated patient experience and yield a good user experience and a majority of plausible answers. Our analysis revealed that GPT-provided answers use either explicit script information or are based on available information, which can be understood as abductive reasoning. Although rare, the GPT-based chatbot provides implausible information in some instances, with the major tendency being socially desirable instead of medically plausible information.

Learning about others: Modeling social inference through ambiguity resolution.

Bayesian accounts of social cognition successfully model the human ability to infer goals and intentions of others on the basis of their behavior. In this paper, we extend this paradigm to the analysis of ambiguity resolution during brief communicative exchanges. In a reference game experimental setup, we observed that participants were able to infer listeners' preferences when analyzing their choice of object given referential ambiguity. Moreover, a subset of speakers was able to strategically choose ambiguous over unambiguous utterances in an epistemic manner, although a different group preferred unambiguous utterances. We show that a modified Rational Speech Act model well-approximates the data of both the inference of listeners' preferences and their utterance choices. In particular, the observed preference inference is modeled by Bayesian inference, which computes posteriors over hypothetical, behavior-influencing inner states of conversation partners-such as their knowledge, beliefs, intentions, or preferences-after observing their utterance-interpretation behavior. Utterance choice is modeled by anticipating social information gain, which we formalize as the expected knowledge change, when choosing a particular utterance and watching the listener's response. Taken together, our results demonstrate how social conversations allow us to (sometimes strategically) learn about each other when observing interpretations of ambiguous utterances.