TABLE 3.
Respondent practical experience of clinical artificial intelligence (AI) over the past year.
| Practice experience of clinical AI | Total (n = 758) N (%) |
Physicians (n = 344) N (%) |
Medical students (n = 414) N (%) |
p-value* |
| Have used decision-support clinical AI systems in practice | <0.001 | |||
| No | 610 (80.47) | 249 (72.38) | 361 (87.20) | |
| Yes | 148 (19.53) | 95 (27.62) | 53 (12.80) | |
| Use frequency** | 0.263 | |||
| Only once a year | 20 (13.51) | 12 (12.63) | 8 (15.09) | |
| At least once every 6 months | 25 (16.89) | 13 (13.68) | 12 (22.64) | |
| At least once a month | 33 (22.30) | 19 (20.00) | 14 (26.42) | |
| At least once a week | 35 (23.65) | 24 (25.26) | 11 (20.75) | |
| Every day | 35 (23.65) | 27 (28.42) | 8 (15.09) | |
| Have met clinical AI error** | 0.207 | |||
| No | 45 (30.41) | 25 (26.32) | 20 (37.74) | |
| Yes | 103 (69.59) | 70 (73.68) | 33 (62.26) | |
| Patient attitudes toward clinical AI** | 0.219 | |||
| Oppose | 2 (1.35) | 1 (1.05) | 1 (1.89) | |
| Neutral | 47 (31.76) | 25 (26.32) | 22 (41.51) | |
| Support | 69 (46.62) | 48 (50.53) | 21 (39.62) | |
| Unclear | 30 (20.27) | 21 (22.11) | 9 (16.98) |
*Chi-square test.
**Only 148 respondents who have used decision-support clinical AI systems in the past year were asked.