The purpose of artificial intelligence (AI) is to help humankind – to help us make more informed decisions faster and more accurately. But when building AI systems, it’s imperative to remember that biases — even unintentional or unconscious — can get introduced into the process.
“AI equity really refers to how effective AI is in a broad range of scenarios with a variety of population groups and demographics,” says Abid Rahman, vice president of innovation at Intouch.
If, for example, a new AI-generated COVID test is 99% accurate when administered to a white Caucasian population but only 40% accurate when administered to South East Asian populations, “millions could potentially become very ill,” says Rahman.
“Likewise, if a drug is developed using genetic data from a certain group of people, and it does not work with the same level of efficacy for other populations, that would be very problematic and potentially even harmful for those other groups.”
Just as we can learn how to move beyond our own biases, we can learn how to build AI systems that avoid biased results. In this piece for our friends at PharmaPhorum, Rahman discusses AI equity and provides practical advice for companies ready to take this important step forward.