Researchers look to deliver ‘unbiased judgment of AI bias’ in medicine

AdobeStock 241101756 1024x576

After an explosion of enthusiasm for the potential of machine learning in medicine, cracks in the foundation emerge.

A growing body of research is focusing on the ways in which medical models can introduce algorithmic biases into healthcare. But in a new paper, machine learning researchers warn that such self-reflection is often ad hoc and incomplete. They argue that to get “an unbiased judgment on AI bias” there needs to be a more routine and robust way to analyze the performance of algorithms. Without a standardized process, researchers will only find the bias they think they are looking for.

Source link

Leave a Comment