Eliminating Bias from AI

Kriti Sharma — a specialist in the development of artificial technologies, delivers a talk on ethical AI that deals with an easily overlooked problem. Aside from the common worry that tech is here to take jobs, the speaker calls attention to an even bigger problem that is flying under the radar.

Algorithms have become an integral part of the decision-making processes of the 21st century. However, AI predicts based on data that is given and more often than not, that information carries with it dangerous human biases that are connected to gender, race, and nationality. As an experienced specialist in the field, Kriti Sharma shares that AI is used to decide who gets a job interview, the rate of one's car insurance, and so on. However, "these decisions are all being filtered through its assumptions about our identity, our race, our gender, [and] our age." As there is a certain "elitism in AI," Sharma argues that we have "reinforced our own bias into the [algorithm]." She gives concrete examples during her talk on ethical AI about the racial and gender profiling that the technology is prone to.

In addition, the speaker calls attention to an obvious gender-based discrepancy between created technology — while assistants like Siri and Alexa are coded female, the male AIs are "more high-powered" referring to IBM Watson, Salesforce Einstein and ROSS, the robot lawyer.

The talk on ethical AI, of course, is hopeful. As Sharma believes that the implementation of smart technology can significantly increase the quality of life, she advocates for diversifying the field and feeding the right values to the algorithm through the diversification of the industry.