Skip to main content
. 2020 Aug 10;22(8):e19104. doi: 10.2196/19104

Table 1.

Comparison of the cost and efficiency of AI systems versus conventional human labor for large-scale screening of COVID-19.

Parameter AIa systems for COVID-19b screening Conventional human labor for COVID-19 screening
Cost May be initially high but significantly decrease with time, leading to lower cumulative cost. May be high or low according to country but will generally have higher cumulative cost, including the added cost of applying preventive measures for direct contact.
Sensitivity (probability of detection) Very high if multiple confirmatory methods are used. Affected by distractors.
Specificity (excluding negative conditions) False positive results may occur due to other conditions having similar signs. When screening is performed by expert health care workers, specificity will be high.
Duration of screening Very short. Relatively long and may require additional employees to shorten the duration.
Number of working hours 24 hours, 7 days a week in addition to the working hours of quarantined physicians and health care workers who are suspected to have COVID-19, who will be able to work from home. No more than 48 hours of work per week and no more than an average of 8 hours of nighttime work per 24 hours of total work [16].
Possibility of COVID-19 infection among examined subjects and examiners Very little or no contact between persons, which significantly decreases the possibility of transmission of COVID-19 infection. High risk of transmission of COVID-19 infection even if preventive measures are followed, as human error may occur.

aAI: artificial intelligence.

bCOVID-19: coronavirus disease.