Table 1 Evaluator participation and rating distribution across scenarios: Each scenario included 11 dialogue lines (6 clinicians, 5 patients)

From: Evaluating LingualAI: a prospective validation of AI-based real-time translation against certified human interpreters

Scenario

Number of evaluators

Clinician dialogue lines

Patient dialogue lines

AI translation ratings

Human interpreter ratings

Total ratings

Scenario 1

8

6

5

1056

1056

2112

Scenario 2

7

6

5

924

924

1848

Scenario 3

4

6

5

528

528

1056

  1. Two AI and two human outputs per line were rated by bilingual clinicians, with totals reflecting combined evaluations for both translation arms.
Back to article page