Content area
Full text
As businesses rely more on AI, changes must be made to reduce bias in the data and results.
I'm not a data scientist. Nor am I an expert on artificial intelligence (AI). I am a documentary filmmaker; as such, I'm in a unique position to ask a lot of questions and interview a lot of individuals who are either experts in their field or experts at just being human. A couple of years ago I set out to make a film about a very human trait: unconscious bias. I was fascinated by the fact that as humans we are inherently biased-it's a survival heuristic we've inherited from our ancestors. But for most of us, the days of having to eye a person or thing that isn't part of our tribe with fearful suspicion are long gone.
Iset out on the streets of cities across the United States asking random subjects, "What is unconscious bias?" and "Are you biased?" Most of the people I spoke with had no idea what unconscious bias meant. When I asked if they were biased, the vast majority said they were not-that they considered themselves "open-minded" and "fair." But when I asked them if they had friends and family members who were biased, every one of them responded with an enthusiastic "yes!"
What this confirmed is what I had been hearing from experts-that we can see bias in others, but not in ourselves.
Is AI Biased?
I began to wonder if artificial intelligence would be the super tool that could solve for human bias. After all, computers aren't human; intelligent as they are, computers don't have human brains. So it stands to reason that computers could take over in situations where humans are unable to be absolutely impartial. Right?
Sadly, no. What I have learned, and what is now increasingly publicized, is that artificial intelligence, in many cases, is also biased.
Humans, even well-intended computer scientists, are...