Gender, Bias, & Artificial intelligence

Whenever I'm speaking about AI at a conference, the question about bias ultimately comes up. You may have seen earlier this year that IBM released the world's largest annotated dataset of facial images to help study bias in facial analysis. The motivation being that to remove bias in AI, we must start with a diverse dataset. However, if we're training a machine to think like a human it's natural to worry that the machine is going to adopt some of the negative traits a person might have shared consciously or unconsciously within that training dataset.

"We want to make sure we eliminate bias. Because Diversity is a big part of innovation." - Amber Grewal, VP Global Talent Acquisition IBM

So when I heard that within IBM we were using Watson for hiring, I wanted to learn more about how we're combating unconscious bias within our AI models to ensure we hire a diverse workforce. I reached out to our in house expert on the topic, @Amber Grewal, Corporate Vice President of Global Talent Acquisition, to learn more. That led to a podcast episode on We Are Only Human with Ben Eubanks.

Here's a sneak peek in this video. 

I want to hear from you how AI can or is combating bias in your solutions, please leave a comment below.