Why does artificial intelligence always seem to get it wrong? This technology has the power to greatly simplify our lives, but instead it magnifies our shortcomings and reflects them back at us. Case in point: Amazon’s hiring algorithm learned, based on successful hires over the past ten years, that women need not apply. How can we program artificial intelligence to think ethically if we can’t even recognize our own shortcomings?
Human bias gets programmed into artificial intelligence algorithms and then the computer isn’t able to distinguish good from bad and ends up going to the extreme with our own biases. In order to ensure that artificial intelligence isn’t programmed with inherent bias, we need to recognize how our own bias plays out in the programming and either eliminate it if possible, or program the AI to ignore the bias if it’s not possible to eliminate it altogether.
AI has gone awry in many ways when it comes to detecting faces of people who are of darker skin tones, and this will start to have serious consequences when it comes to driverless cars not being able to see certain pedestrians or facial recognition software misidentifying certain groups of people.
Learn more about bias in artificial intelligence from the infographic below.
Comments
Log in or sign up to join the conversation.