“The Myth of Objectivity” - How AI Inherited Human Bias

Abstract

Artificial intelligence algorithms, created with the intention of learning and performing complex tasks - the AI algorithms in short, were supposed to reduce human prejudice affecting decision-making processes. Yet despite the colossal efforts they still inherited something after their creators – the bias. Since humans as a species are prone to favoritism so are their creations. The origins of this phenomenon are to be found not only in the creators’ personal beliefs and cultural and societal conditions commonly associated to exposure to outer sociological factors, but also in the developed technology itself. This work describes the potential sources of bias and a study conveyed on a group of independent participants who were asked to answer a questionnaire which results were later used to train an AI model in terms of bias detection.