AI in Cancer Diagnosis Faces Bias and Unexpected Insights
Originally Published 26 days ago — by SciTechDaily
A study reveals that AI models used for cancer diagnosis from pathology slides can inadvertently read patient demographics, leading to biased results across race, gender, and age groups. Researchers developed a new framework called FAIR-Path that significantly reduces these biases, highlighting the importance of testing and improving fairness in medical AI tools to ensure equitable cancer care for all populations.