
Why grading AI is still hard, according to Hugging Face
Hugging Face policy chief Irene Solaiman said putting testers inside AI companies won't easily fix how AI is graded, speaking at TechCrunch Disrupt on Oct. 19. There are no agreed rules for who qualifies as a tester, she said, and testing AI is still a new science with no reliable way to compare scores, like exams with different grading.
Published