Table 6 Score distribution of chatbot responses according to the modified DISCERN scale and GQS classification
From: Can artificial intelligence models serve as patient information consultants in orthodontics?
ChatGPT-3.5 | ChatGPT-4 | Gemini | Copilot | ||
|---|---|---|---|---|---|
n (%) | n (%) | n (%) | n (%) | ||
Modified DISCERN Score | Poor (8–15 point) | 3 (15%) | 1 (5%) | 0 (0%) | 0 (0%) |
Moderate (16–31 point) | 17 (85%) | 19 (95%) | 19 (95%) | 19 (95%) | |
Good (32–40 point) | 0 (0%) | 0 (0%) | 1 (0%) | 1 (5%) | |
GQS Score | Low quality (Score 1 or 2) | 0 (0%) | 0 (0%) | 0 (0%) | 0 (0%) |
Moderate quality (Score 3) | 11 (55%) | 7 (35%) | 4 (20%) | 2 (10%) | |
High quality (Score 4 or 5) | 9 (45%) | 13 (65%) | 16 (80%) | 18 (90%) |