Artificial intelligence isn't just science fiction any more. It's all over the place, and it has the potential to change the way we do everything from harvest crops to conduct wars. However, if your image of an artificially intelligent machine is one that thinks for itself and is capable of perfect, rational thinking without any biases, think again. Evidence is increasingly suggesting that the artificial intelligence currently being built is deeply vulnerable to the biases of the people and information programming them. New research from the University of Virginia found that artificial intelligence can exhibit sexist behaviors, even if programmers don't intentionally instill those values. Considering how much the future of many industries (not to mention our personal lives) may depend on various forms of AI, bias of all kinds is a serious issue.
Data released recently suggests that the global impact of AI will be around $15 trillion — yes, trillion with a T. Analysts think robots will influence everything from health and medicine to manufacturing and the financial sector. And with the first ever album composed by an AI in collaboration with a human is also being released this week, the sky's the limit. But looking at how AIs learn about the world, and figuring out what they do with that knowledge, has revealed that artificial intelligence may just reflect the same biases and prejudices that their creators exhibit — namely, sexism and racism.