Artificial intelligence programs that diagnose disease by analyzing images still require human judgment
Earlier this month a team of computer scientists at Stanford University announced that their artificial intelligence program had diagnosed pneumonia on chest X-rays more accurately than human radiologists.
“Should radiologists be worried about their jobs?” one of the scientists, Andrew Ng,tweeted.
Probably not. In fact, they ought to be relieved: AI is more likely to make their jobs easier than redundant.
Radiologists today face the same 21st century problem of most professionals: too much data. Machine learning, the latest iteration of AI, can sift through that data and spot patterns humans might miss. But machines don’t know what to do with that information—at least not yet. Only humans do.
