Google DeepMind AI Outperforms Doctors in Breast Cancer Screening Tests
A new Google Health AI system shows greater accuracy than human radiologists in detecting breast cancer, significantly reducing both false positives and false negatives.
Artificial intelligence from Google's DeepMind demonstrates superior accuracy in spotting breast cancer compared to human doctors. Researchers from Imperial College London and Google Health train the computer system using X-ray images from nearly 29,000 women to identify subtle abnormalities.
The AI system achieves significant improvements in diagnostic accuracy across both U.K. and U.S. studies. In the United States, the technology reduces false positives by 5.7% and false negatives by 9.4%, outperforming all six human radiologists in an independent comparison.
This breakthrough offers a potential solution to severe radiologist shortages within the U.K.'s National Health Service. Developers expect the system to become even more effective once it integrates supplementary patient history, providing a vital tool to alleviate backlogs and improve patient care.