Diversity and Inclusion in AI Development Teams
According to McKinsey, companies that are gender diverse are 15% more likely to have financial returns above their respective national industry medians. Companies with ethnically diverse workforces experienced 35% higher returns than other companies in their respective industries. And according to researchers at Bain & Company, Fortune 100 companies with three or more women on their board had a 21 percent higher return on equity than other corporations.
It is essential to recognize that artificial intelligence will have an increasingly important role in global business. As AI systems grow more sophisticated, they can handle increasingly complex tasks—everything from basic data collection and pattern analysis to more advanced processes like market forecasting. However, there are some things computers just can’t do as well as humans yet. And one of those things is to deal with ethics. While we might hope our machines would operate ethically at all times, we know full well they don’t – and so does everyone else.
AI can potentially change the face of business as we know it, and those changes will occur in our top leadership positions first. That’s why diversity and inclusion should be at the forefront of every company’s strategic plan for AI adoption and implementation moving forward. AI ethics discussions need to include diverse leadership perspectives and voices to ensure that diverse and inclusive voices are heard and considered in forming ethical AI guidelines and best practices. So if you want your business to stay ahead of AI as its capabilities expand, start by ensuring you have diverse leaders at every level.
It’s no surprise that investors are seeking diversity in leadership roles. But why? What is it about diverse leadership teams that correlate to better company performance? The reasons may be surprising. It turns out there’s more to diversity than meets the eye.
First, creating an inclusive culture where everyone feels comfortable bringing their authentic selves to work each day allows employees to bring different perspectives on problem-solving. While perspectives are often regarded as something positive, it’s also important not to mistake multiple perspectives for differing viewpoints. Multiple perspectives are valuable because they mean people share different experiences based on variables like age, education level, social class background, etc. In contrast, differing viewpoints suggest that people genuinely think differently. Different backgrounds give people an informed take when brainstorming solutions.
Diversity is a risk-mitigation strategy. By cultivating unique perspectives from different groups, companies can improve decision-making across their business at large.
Additionally, socialization into varying cultures makes room for understanding and appreciation of different ways to do things. Because we interact primarily with those who look like us or hold similar opinions (for good reason), we tend to approach tasks and experiences very similarly. This can negatively impact our ability to achieve goals because we fail to identify alternate paths toward solutions – ultimately limiting our innovation potential. We often become blinded by our own behaviors — even when they inhibit progress — before we even realize what’s happening.
For example, a woman may steer away from innovation suggestions simply because she didn’t recognize her male peers’ ideas were innovative. She’ll question whether she’s being too critical of their ideas without realizing that feedback plays a crucial role in developing new concepts. Or someone who grew up thinking women shouldn’t lead will struggle to find any value-add he receives from female leaders, regardless of what he says publicly. Such biases are hard to overcome unless all types of people are present within any given organization. That way, individual characteristics can provide benefits rather than hindrances, leading organizations to greater success over time.
Finally, with AI systems making decisions that impact our lives—such as who to hire, who to detain, who gets insurance benefits—we need diverse people creating them. This isn’t because there is a bias against minorities or women in algorithms, but rather because we have limited data on minority populations. Put another way: When all you have is a hammer, everything looks like a nail. And if AI development follows its current path, society will end up with only one tool at its disposal for complex problems. Even if AI does learn from diversity, it can take several years for AI technologies to mature from concept to implementation across an organization. In other words, change will not happen overnight—but now is definitely the time to think about it.
Fostering an open, forward-thinking environment that embraces diversity and inclusion begins with looking critically at your hiring process. If it’s culturally biased, your entire team is going to skew in one direction. Part of promoting diversity and inclusion involves making sure that hiring managers are aware of their own biases. It’s easy to say this person has a great attitude or they would really fit in here, but it may reflect different cultural attributes that actually make them ill-suited for success.
You need to separate perception from reality – not just how others see you but also how you see yourself. It sounds simple, but it takes work to get past these assumptions. And once hires are made, research shows that failing to acknowledge different viewpoints hinders progress. It’s not enough to hire a diverse group of people. Real progress requires accepting diversity and actively encouraging it through socialization opportunities, sponsorships, and executive-level commitment.
Diversity isn’t a goal or an end in itself. It’s simply an avenue to innovation and success that we ignore at our own peril. Not only that, but it’s the right thing to do.
Hiring diverse teams is one of the steps we can take to recognize and eliminate AI bias. According to a study done by researchers at Stanford University, Google, and New York University, commercial datasets used for machine learning often reflect the prejudices and preferences of their designers. These biases come from language and accents: if a person is not American or speaks with an accent, then their voice will likely be left out. Even search algorithms discriminate against women because they tend to ask questions that require shorter answers, favoring white men who have more spare time on their hands to do research. In addition, women’s queries tend to contain superlatives such as best and greatest, which also feed into these algorithms.
Artificial intelligence (AI) has already changed how we live, work, and play. But AI can also be used to drive diversity efforts, helping make room for more women, people of color, LGBTQ individuals—and other historically underrepresented groups—in top leadership positions.
Some AI tools make it possible for organizations to vet candidates based on set criteria such as age, education level, professional experience, skills, etc. Using these tools leaves less room for unconscious bias or discrimination against candidates based on factors like gender or ethnicity. Hiring managers can review AI-vetted candidates side by side with traditional resumes so they can base hiring decisions on fit and skills rather than name or looks. Employers looking to improve diversity and inclusion in hiring—or promote diverse employees within existing teams—can integrate such tools into existing HR systems and processes so everyone benefits from AI technology right away.
Diversity is not just a feel-good talking point; it’s proven to improve businesses by making them smarter, more innovative, better run—and more profitable. This means any company that wants to succeed should embrace diversity and inclusion as core values: The winners will be those who find ways to maximize their talent and truly reflect their customers.
It is critical that organizations define their vision for how AI fits into their long-term strategy. But more importantly, they must also explicitly state how they want their AI ethics to guide them today.
Maria Mac Andrew is the head of Diversity and Communities at AI Ethics world and the Co-Founder of a Global Ethical AI Foundation.
At AI ethics World, we help businesses and communities tackle these big questions by introducing our expertise in diversity leadership strategies and training our clients to apply AI ethics into practice.