Table 1.

Accuracy of consistency to ground truth of main and differential diagnosis between the chatbots and ground truth

	Chatbot	Consistency to ground truth (ratio, percentage, 95% CI)
Main diagnosis matched	GIA-CB	39/50, 78% (0.67–0.96)
Main diagnosis matched	g-CB	27/50, 54% (0.40–0.68)
Main diagnosis contained	GIA-CB	45/50, 90% (0.82–0.98)
Main diagnosis contained	g-CB	37/50, 74% (0.62–0.86)
At least one fitting differential diagnosis	GIA-CB	49/50, 98% (0.94–1.00)
At least one fitting differential diagnosis	g-CB	46/50, 92% (0.85–1.00)

CI Confidence interval, GIA-CB Gastrointestinal imaging-aware chatbot, g-CB Generic chatbot