Artificial Intelligence in the Grip of “Diversity Crisis”


Researches at the AI Now Institute found out that as the technology becomes an industry, the artificial intelligence industry is said to lack diversity as it is dominated by white male scientists / Photo by: Halfpoint via Shutterstock


Too white and too male—this is the description being used to for the wider artificial intelligence industry, as researchers from the AI Now Institute found in a research that this overwhelmingly unfair ratio of race in the industry could potentially raise questions about where the industry is headed.

So why do we need diversity in the industry? According to the researchers in their study, as reported by The Verge, most of the people that are underrepresented in the AI workplace are women and people of color, as data reveals that 80 percent of the AI industry is actually made up of white males. Even though they are skilled workers, people of color are also more likely to be “sidelined.”

This is not a good sign if we want to make artificial intelligence (AI) inclusive of every experience and perspective. For instance, AI helps people in Ghana with identifying crop health through innovations like machine learning and AI, and that’s a great example of diversity entering the industry and producing helpful AI technology that aligns with perspectives and the situations of minorities.

If there are only white, male workers in the AI industry, this results in a narrow perspective of the possibilities that lie in artificial intelligence.

The other ways in which a biased perspective can hurt minorities in the future of AI are technologies like facial recognition, which “can disproportionately affect historically marginalized groups.”

What’s worse is that even though there have been “steps” supposedly taken in consideration to fix the problem of diversity in the AI industry, none has yielded enough significant results for it to be reflected.

To truly amend this, the researchers suggested firstly that companies should honor and improve transparency “by publishing more data on compensation, broken down by race and gender, and by publishing harassment and discrimination transparency reports.”