From: Exploring the potential of ChatGPT as an adjunct for generating diagnosis based on chief complaint and cone beam CT radiologic findings
Number of answers (percentage)
Number of mistakes (percentage)
Without text error
346 (88.7%)
0
Text error
Imprecise language
44 (11.3%)
28 (63.6%)
Hallucination
13 (29.5%)
Wrong location
3 (6.8%)