Gender bias and stereotyping in human interaction is not uncommon. For example, consider the following description of a man and woman performing very similar tasks. “She is a bossy and assertive individual”. Contrast that with “He is a dynamic and confident person”. Indeed, this could perhaps be representative of individual traits. However, when similar behavioral descriptors often repeat in social media, it subliminally suggests “conformity into stereotypical societal roles” as expected and approved behavior. Such perceptions not only impede societal progress and evolution but also handicap individual growth.

Using novel statistical metrics, this paper first demonstrates the strong positive connotation between gender and expected “societal roles” in social media. Next, Artificial Intelligence (AI) algorithms are developed to mathematically quantify the bias. The fundamental technique of applying AI algorithms is motivated by the popular game of “Hide-and-Seek” and is largely language independent. Subsequently, in the same spirit as the popular tool “SpellCheck” (which automatically corrects for typographical errors), an AI based software called “BiasCheck” is developed for detecting this biased language. Results substantiate the presence of male and female gender stereotypes and also highlight differences between male/female societal behavioral perceptions. In particular, female gender stereotypes are shown to be relatively easy to determine even with simple techniques. This inter-disciplinary research readily extends to studying any form of bias and is an important advancement in the algorithmic study of human behavior and prejudic e in social media as well as a novel application of AI for “social benefit and good”.