Table 2.
Correlation among constructs.
| Anxiety (PRA) | Biases (PSB) | Privacy (PPC) | Trust (PMT) | Communication barrier (PCB) | Unregulated (PUS) | Liability (PL) | Risks (PR) | Benefits (PB) | |
|---|---|---|---|---|---|---|---|---|---|
| Biases (PSB) | 0.595 | — | |||||||
| Privacy (PPC) | 0.506 | 0.417 | — | ||||||
| Trust (PMT) | (−0.656) | (−0.494) | (−0.463) | — | |||||
| Communication barrier (PCB) | 0.569 | 0.458 | 0.470 | (−0.528) | — | ||||
| Unregulated (PUS) | 0.597 | 0.545 | 0.544 | (−0.515) | 0.525 | — | |||
| Liability (PL) | 0.587 | 0.490 | 0.529 | (−0.566) | 0.634 | 0.688 | — | ||
| Risks (PR) | 0.726 | 0.570 | 0.535 | (−0.756) | 0.622 | 0.587 | 0.639 | — | |
| Benefits (PB) | (−0.591) | (−0.442) | (−0.426) | 0.679 | (−0.539) | (−0.403) | (−0.480) | (−0.620) | — |
| Intention to use (INT) | (−0.634) | (−0.479) | (−0.459) | 0.777 | (−0.607) | (−0.477) | (−0.562) | (−0.718) | 0.825 |
All correlations reported in this table are significant at P < .001.
Positive numbers would indicate similar responses, and negative numbers indicate that as 1 construct mean goes up, the other goes down.
For example, PCB vs PB = −0.539, indicating that as ratings increase for PCB, they tend to decline for PB.
Thus, if responding strongly agree (5) to I am concerned that AI tools may eliminate the contact between healthcare professionals and patients, respondent may have also responded strongly disagree (1) to I believe AI-based services can improve diagnostics.
Abbreviations: AI, artificial intelligence; PSB, perceived social biases; PPC, perceived privacy concerns; PMT, perceived mistrust in AI mechanisms; PCB, perceived communication barriers; PUS, perceived unregulated standards; PL, perceived liability issues; PR, perceived risks; PB, perceived benefits.