Skin cancer can be deadly. California residents have heard all the messaging about covering up and using sunscreen. But even among people who take precautions, skin cancers can occur. In fact, the rates of melanoma have been going up for three decades. Increasingly, medicine is turning to new diagnostic tools, including artificial intelligence. However, these tools are far from perfect.
Problems with AI diagnoses
One of the biggest problems when it comes to the use of AI in dermatology is related to race. AI detects skin cancers in light-skinned subjects much more accurately than for darker-skinned subjects. This is a real problem for African Americans, who already face higher mortality rates from melanoma.
The issue is that the data that “teaches” the AI how to diagnose skin cancers comes mostly from images of light-skinned people. It’s important to incorporate more diverse skin tones in these datasets. Otherwise, some of the most high-risk patients will remain underserved.
Misdiagnosis is a big problem
Many clinicians are not properly trained in identifying skin cancers in people of color. AI faces the same issue. AI sometimes misidentifies a spot as cancerous when it isn’t. At other times, it misses melanomas. Cancer misdiagnosis is a serious problem, especially with something as deadly as melanoma. Some African Americans in their 20s have died due to these discrepancies.
If you or a loved one has been misdiagnosed in a skin cancer case, it’s important to seek legal advice. A skilled attorney may be able to help you pursue a civil suit. They may be able to help you determine who to sue, and for what amount. Damages will not make the situation perfect, but they can help to cover things like medical bills and lost wages.