DeepMind is a British Artificial Intelligence company that was acquired by Google in 2014. It has turned its attention towards breast cancer detection, a field that is often faced with problems regarding false alarms, where a mammogram is sometimes judged as abnormal even when there is no cancer present.
This can lead to unnecessary distress and medical treatments that could pile up into jaw-dropping bills. A study in 2015 estimated that these false alarms had cost the US over $5 billion a year.
Fortunately, DeepMind has developed a new AI model for breast cancer detection that can beat experts by minimizing the likelihood of incorrect readings.
The company trained this AI using anonymous data taken from numerous patients across the UK and the US and the results were fascinating. It was able to reduce false readings by up to 9.4% and the decrease in incorrect screenings was less in the UK compared to the US, suggesting that the British breast cancer-detecting systems are more accurate than their American counterparts.
Human experts make use of patient history and mammograms from different time periods to judge the presence of cancer. DeepMind’s AI was able to make more accurate readings even when it only had access to the most recent mammograms.
However, the developers behind this AI claim that it is still in its early stages and a lot of support from healthcare providers is required before it can be made available to the public.
Still, it’s good to see AI making headway in medical care and diagnosis.
Surely this is just down to humans actually caring for people vs Ai. A human might be sure that there’s no cancer but wants to test because it would be horrible for the patient if they miss it. While and Ai wouldnt care about that and would just use probabilities. But surely the human system is better incase it does find something.