As a Spanish woman, I used to meet my friends for dinner at 10pm. I thought that was the perfect time to make the most out of a day. Fast forward to today, after living in Amsterdam, working in an international environment and settling down in Germany, I would never organize a dinner so late. I firmly believe it makes more sense to eat earlier.

We all carry with us a wealth of life experience, a kind of book of stories, which we consult to make sense of the world. Our identity (gender, race, age, origin), neurological differences, our studies and our experiences influence the content of that book directly, and therefore the way we reason.

I am still a woman of Spanish origin, but my reasoning when I organize dinner has changed. It makes so much more sense to have a light lunch at midday and dinner in the early evening. That is now part of my book.

Diversity is not a recipe for success for every team. If we need to carry bricks, we need the strongest individuals. But if we need to solve complex problems, it is proven that diversity helps. McKinsey published findings in 2018 that companies in the top-quartile for ethnic/cultural diversity on executive teams were 33% more likely to have industry-leading profitability.

The reason for this advantage is not strictly the variety of origin, ethnic or gender. It is because, due to those differences, team members have different chapters in their books. I am an engineer, and I have lived different experiences as a woman, and as a Spanish person. These stories are in my book and influence how I see the world. Diverse people reason differently and can come up with more and new ideas. It is a matter of cognitive diversity.

Data and AI-driven organizations and cognitive diversity

Gartner acknowledges that data and analytics are the critical elements needed to accelerate an organization’s digitization and transformation efforts, which are essential to compete in the digital economy.

Why is cognitive diversity a top success factor for your data-driven goals and AI plans in your organization? Data is giving us information for making informed decisions. But the reality is that data can be interpreted depending on who reads that data.

One of the best examples is that human fertilization has traditionally been explained as a kind of fairy tale by male scientists, with competitive sperms fighting for a passive female egg, even though this is not the case, as has been highlighted by more diverse scientific teams.

Artificial Intelligence interpretation of data can be as biased as the human intelligence behind it. What can happen when you have an excellent AI team but it is not cognitively diverse? Amazon’s recruiting algorithm was gender-biased, discarding competitive female candidates for roles after learning that the percentage of women in those positions was lower.

When Youtube launched a video upload app for iOS, between 5 and 10% of videos uploaded by users were upside-down because the design team was right-handed, so they had not taken into account the way left-handed people manage a phone. Another example was when AI from Google Photos labelled African Americans as gorillas. It is still raised in shareholder meetings, as a clear demonstration of why Google needs to become more diverse.

These are examples of bias in AI products. Could you imagine what would happen if we are planning an AI-driven organization, with AI-based decisions taken across functions? What is the way to avoid that inherent bias?

The answer is cognitive diversity. Non-homogeneous teams are more capable than homogenous teams of recognizing their biases and solving issues when interpreting data, testing solutions or making decisions.

How to boost cognitive diversity in your workforce

The more content in our books to contrast and consult, the more ideas and perspectives we have. It is an advantage when we add new chapters, combine different books or increase their number. Similarly, there are three ways to ensure your business boosts cognitive diversity:

1) Leveraging your current teams

Gartner said that by 2020, 50% of organizations will lack sufficient AI and data literacy skills to achieve business value. You need to ensure that your current workforce is ready for the new demanding data and digital skills. From HR to Sales, Marketing, Finance, IT, all your organization needs to understand the language of data. It is a win for all: each person can grow in their current roles, or be eligible for new opportunities, while increasing the value of the business.

2) Making data scientists your business partners

While we are used to roles like HR Business Partners or Finance Business Partners in the organization, the value of a true business partnership with Data Scientists is often overlooked. Combining the reasoning of data scientists with the understanding and experience of the business functions results in decision-making processes with measurable and data-driven evidence, and new opportunities and challenges within and outside an organization. Businesses need a diverse group of citizen data scientists to explore hypotheses and insights.

3) Attracting and retaining diverse talent

Vision and strategy need to include diversity. Focus on selection practices, for example, by offering training for selection procedures and impartial interviewing to your recruiters. Create a baseline with metrics of the current diversity numbers and audit the evolution.

A way to attract candidates is to create strategic partnerships that link with diverse data science talent pipelines, such as Women in Big Data. Define programmes and initiatives for under-represented groups, like many big corporations are already doing.

Conclusion

Big data and Machine Learning are providing exciting opportunities for businesses and their workforce. But there is a high risk of failure if we do not handle the potential bias that surrounds them. Cognitive diversity is the way to ensure we get the best business value out of artificial intelligence. All you need is a diverse workforce, who all consult different books when sharing their ideas. Are you ready?