Smart machines and Artificial Intelligence (AI) are set to dominate our future, changing every aspect of our lives. Already it determines our behaviours, thought processes, purchasing habits and worldviews.

From facial recognition to self-driving cars, AI and machine learning are set to become intrinsically interwoven into the fabric of daily life and in years to come it will increasingly shape our perception of the world around us.

Companies and governments are welcoming these new technological trends as they have the potential to let computers make decisions and take action. For example, already we are seeing AI and machine learning being deployed in fields such as policing, healthcare, academia and the judiciary system.

Of course, they are only as good as the data sets used to train them; and this is wherein the problem lies. AI researchers are predominantly white or Asian males and as a result, other groups are not being reflected in the data. In particular, women are severely underrepresented in this area, which has resulted in gender-biased AI.

Research conducted by WIRED and Montreal-based startup, Element AI, revealed that only 12% of leading machine learning researchers were women. This gender imbalance suggests that the group spearheading this technological revolution is less inclusive than the broader tech industry. These figures are, however, not surprising considering the scarcity of women in the field.

Earlier this year MIT reported that only 20% of engineers at Google and Facebook were women, despite their efforts to hire more. Hanna Wallach, an AI researcher and co-founder of the Women in Machine Learning Conference, said that only 13.5% of those working in machine learning are female.

Diversity Brings a Much Needed Fresh Perspective

Those arguing for greater diversity in AI believe that failing to do so will increase the risk that AI systems could have harmful effects on the population. Anima Anandkumar, a professor at the California Institute of Technology who previously worked on AI at Amazon, asserts that “diverse AI teams are more likely to flag problems that could have negative social consequences before a product has been launched”.

Already there have been instances of AI teams releasing biased data and systems into the world. For example, researchers at the universities of Virginia and Washington showed that two large image collections used in machine learning research, including one backed by Microsoft and Facebook, teach algorithms a skewed view of gender. Images of people shopping and washing are mostly linked to women, for example.

Boston University discovered that the word programmer was more likely to be linked to a male than a female. Along with redressing the gender imbalance, the AI community needs better representation of ethnic minorities. Earlier this year researchers from MIT’s Media Lab and Microsoft found that facial analysis services that IBM and Microsoft offered to businesses revealed that facial recognition algorithms were 12% more likely to misidentify dark-skinned males.

The companies’ algorithms were near perfect at identifying the gender of men with lighter skin but frequently erred when presented with photos of women with dark skin. This flaw had been on the market for more than a year when it was discovered.

Both Microsoft and Facebook are already developing bias detection tools, which will automatically issue an alert of an algorithm is making an unfair judgment about a person based on race, gender or age. As we become more dependant on AI and machine learning for decision making this imbalance must be addressed or women and minority groups could face even worse discrimination in the future.