Table 1.
Summary of ChatGPT’s performance on MedScape clinical case challenges.
Fig 1.
Percentage of correct answers, most common answer and correct answer despite the majority incorrect by ChatGPT 3.5 with MedScape clinical case challenges.
Fig 2.
Confusion matrix evaluating the diagnostic accuracy of ChatGPT 3.5, considering each answer within the 150 MedScape clinical case challenges.
Fig 3.
Receiver Operator Curve (ROC) for the diagnostic accuracy of ChatGPT 3.5 answers within 150 MedScape clinical case challenges.
Fig 4.
Cognitive load of ChatGPT 3.5 answers given in response to 150 MedScape clinical case challenges.
Fig 5.
Quality of medical answers given by ChatGPT 3.5 response to 150 MedScape clinical case challenges.
Table 2.
Qualitative analysis of the strengths associated with ChatGPT’s answers in response to MedScape clinical case challenges.
Table 3.
Qualitative analysis and examples of the weaknesses associated with ChatGPT’s answers in response to MedScape clinical case challenges.