AI has taken on an almost-science fiction sheen, says Jiaying Claire Wu, alumna of Columbia’s Applied Analytics M.S. program. In a TEDxColumbiaUniversity Talk earlier this month, Wu challenged viewers to rethink what they envision when they talk about AI and build from there to better understand what the AI is evaluating when it makes a prediction. This shift in perspective, she says, can build general trust in AI and its potential to benefit and streamline our lives in broad areas from customer service to healthcare.
“Ask what drives the model of predictions, and whether those drivers make sense to you,” says Wu, a Data Scientist for consulting firm Deloitte. When data scientists are considering the accuracy of AI-based predictions, instead of taking on faith that the machine knows best or getting frustrated when an incorrect prediction is made, they should take a look at how the AI is coming to those conclusions.