Attitudes, perceived benefits and risks | N (%) |
Attitude towards AI-supported mammography screening (Q9) | (N=47) |
Positive | 11 (23.4) |
Somewhat positive | 27 (57.4) |
Uncertain | 6 (12.8) |
Somewhat negative | 1 (2.1) |
Negative | 2 (4.3) |
Preferred use of AI in mammography screening (Q13) | (N=47) |
AI as triage tool | 6 (12.8) |
AI as stand-alone reader | 2 (4.3) |
AI as replacement of one in double reading | 21 (44.7) |
AI as addition to double reading | 14 (29.8) |
Not at all | 4 (8.5) |
Potential benefits of AI-supported screening (Q10) | (N=47) |
To a high degree | 13 (27.7) |
To a somewhat high degree | 24 (51.1) |
Uncertain | 6 (12.8) |
To a somewhat low degree | 2 (4.3) |
To a low degree | 2 (4.3) |
Potential risks of AI-supported screening (Q11) | (N=47) |
To a high degree | 6 (12.8) |
To a somewhat high degree | 10 (21.3) |
Uncertain | 16 (34.0) |
To a somewhat low degree | 14 (29.8) |
To a low degree | 1 (2.1) |
Perceived risk of overconfidence in AI assessments (Q15) | (N=47) |
To a high degree | 4 (8.5) |
To a somewhat high degree | 9 (19.1) |
Uncertain | 20 (42.6) |
To a somewhat low degree | 13 (27.7) |
To a low degree | 1 (2.1) |
Perceived risk of non-representative training data (Q38) | (N=47) |
To a high degree | 2 (4.3) |
To a somewhat high degree | 9 (19.1) |
Uncertain | 24 (51.1) |
To a somewhat low degree | 9 (19.1) |
To a low degree | 3 (6.4) |
Perceived risk of inferior AI performance on certain risk groups or specific type of cases (Q39) | (N=47) |
To a high degree | 6 (12.8) |
To a somewhat high degree | 13 (27.7) |
Uncertain | 22 (46.8) |
To a somewhat low degree | 5 (10.6) |
To a low degree | 1 (2.1) |
AI, artificial intelligence.