Table 10.
|
Estimate of the meana (95% CI) | Design effect | ||
C01b —machine reads medical test, diagnoses, and recommends treatment |
|
|||
|
Accuracy | 1.49 (1.46-1.53) | 1.98 | |
|
Human contact | 1.78 (1.74-1.81) | 1.95 | |
|
Responsibility | 1.78 (1.75-1.82) | 1.98 | |
|
Explanation | 1.86 (1.82-1.90) | 1.96 | |
|
Fairness | 1.87 (1.83-1.91) | 1.91 | |
|
Speed | 2.08 (2.04-2.12) | 1.88 | |
|
Reducing costs | 2.30 (2.25-2.34) | 1.92 | |
C02—machine triages when you are unwell |
|
|||
|
Accuracy | 1.56 (1.51-1.61) | 1.73 | |
|
Responsibility | 1.76 (1.71-1.81) | 1.75 | |
|
Human contact | 1.81 (1.75-1.86) | 1.72 | |
|
Explanation | 1.87 (1.82-1.93) | 1.76 | |
|
Speed | 1.90 (1.85-1.95) | 1.64 | |
|
Fairness | 1.94 (1.88-2.00) | 1.81 | |
|
Reducing costs | 2.43 (2.36-2.50) | 1.74 | |
D01—machine processes application for unemployment benefits (data sharing required) |
|
|||
|
Accuracy | 1.61 (1.56-1.65) | 1.53 | |
|
Fairness | 1.80 (1.75-1.85) | 1.56 | |
|
Explanation | 1.86 (1.80-1.91) | 1.61 | |
|
Personal tailoring | 1.87 (1.82-1.92) | 1.58 | |
|
Human contact | 1.88 (1.82-1.93) | 1.54 | |
|
Speed | 1.99 (1.93-2.04) | 1.58 | |
|
Reducing costs | 2.51 (2.45-2.58) | 1.59 | |
D02—chatbot advises about carer payments |
|
|||
|
Accuracy | 1.60 (1.55-1.64) | 1.6 | |
|
Fairness | 1.81 (1.76-1.87) | 1.68 | |
|
Personal tailoring | 1.82 (1.77-1.87) | 1.67 | |
|
Human contact | 1.83 (1.77-1.88) | 1.63 | |
|
Speed | 1.91 (1.86-1.97) | 1.71 | |
|
Explanation | 2.02 (1.96-2.08) | 1.72 | |
|
Reducing costs | 2.60 (2.54-2.67) | 1.71 |
aMeans and CIs adjusted for weighting.
bCode (eg, C01) indicates question number in instrument.