Why We Can’t Move Forward Without Diversity in AI

From navigation apps to facial recognition, smart assistants, autonomous vehicles, and everything in between, artificial intelligence is ingrained into our everyday lives.

The artificial intelligence market is rapidly growing, but there is still a lack of diversity, which is a cause for major concern.

Diversity allows for a difference in viewpoints to help optimize AI development and reduce biases that might come from the data and the criteria that is being inputted. In addition, a diverse workforce will help decrease the likelihood of discrimination by these systems.

Join Angle and Quyen as they dive into why we can’t move forward without diversity in AI, how to attract and retain diverse talent, different biases in AI and how we can work to combat them, and much more.

About Angle

“Artificial Intelligence is the Fourth Industrial Revolution… Well, you can’t have a revolution without Black Women.”

Angle Bush | Founder, Black Women in Artificial Intelligence

After attending an event that did not fully reflect the diversity needed to ensure access and opportunity in the artificial intelligence sector, Angle decided to do something about it. In 2020, she founded Black Women in Artificial Intelligence, an organization whose mission is to Educate, Engage, Embrace, and Empower Black women in the Artificial Intelligence industry.

In less than 2 years, the organization has seen phenomenal growth, becoming a global organization whose members span four continents. In 2021, Angle spearheaded the campaign to create National Black Women in Artificial Intelligence Day, which is now celebrated annually on August 1 to amplify the voices of Black women and honor their contributions to Artificial Intelligence.

Aligning with the mission to provide access and opportunity, the organization has also developed partnerships with some of the top companies in the nation, including Amazon, Capital One, Expedia, and NVIDIA. From learning opportunities to mentorship programs, the organization continues to create initiatives that focus on creating spaces for Black women to thrive.

Before starting Black Women in Artificial Intelligence, Angle worked in the oil and gas industry for over fifteen years serving in several positions from Process Improvement Manager to Vendor Relationship Manager. She also served as the Executive Producer of Your15Minutes Radio, an online podcast that focuses on entrepreneurship. All of which have led her to this mission, this moment.

Angle earned a Master’s degree in Administration and a Bachelor of Science degree in Community Development with a concentration in Public Administration from Central Michigan University. She also attended Harvard University, where she studied History.

About Quyen

Being the Vice President of Swoon Consulting allows Quyen to merge her passion for data and leadership. Through this position, she can help clients solve data and analytics challenges driven by globally diverse industry experts. In addition, she is an advocate on Swoon’s DE+I committee and a member of the Forbes Business Development Council.

Quyen immigrated to the United States post-Vietnam War with her family as a child. That experience and seeing how hard her family worked to get where they are today pushed her to take risks and every opportunity available to her. Essentially, her family risked everything to get them there, which was always very powerful to her. She believes that where you come from, your upbringing and the things you have been through all add a unique perspective to who you are and who you become as a leader.

Megan: Why is diversity so crucial in the development of AI?

Angle: When we think about diversity and artificial intelligence, there have been so many things in STEM and tech. It is essential that when we create these technologies, we make sure that there is diversity in the room. There is a diversity of intellect, abilities, and skills in the room. We want to ensure that there is no harm done to communities because, unfortunately, that is currently happening in AI within specific areas.

Quyen: There was an article published on Forbes in 2020 on why it is so important to have diversity in AI because it can be pretty dangerous if we do not. There could be flawed AI systems that perpetuate gender or racial biases if we do not have the correct data and are not assessing that data correctly. So, it is not only essential, but it is a must that there is diversity in AI. Angle, I especially love what you are doing with Black Women in AI because it brings together this diversity that we need to ensure that these disasters do not happen in the future.

Angle: It is not just about being in the room, and we have to be aware of that. We can be in the room and not have a voice or a vote, so we have to make sure that those in the room have the power and the authority to say, “hold on, let’s stop and reassess whatever technology we are trying to create.” Diversity should not be the last thing but the first thing we have a conversation about.

Megan: Being that there is currently a lack of diversity within the AI industry, how can we work to support diverse talent pursuing a career in AI?

Angle: The best way to support diverse talent is to recruit them. Let there be authenticity and sincerity in the recruitment process. When it comes to partnerships, one of the things we talk about in Black Women in AI is we set the initial meeting, and we tell them that we need this partnership not just to be transactional. We want it to be transformational. The same sentiment has to go along with recruiting diverse talent or trying to hire diverse talent, that we can hire 100 diverse candidates, and they can all leave within the next six months. We have to question if we were successful at that point and go back to reevaluate the corporate culture. Is the corporate culture open enough to support diverse talent in terms of promotions, having a seat at the table, recognizing their abilities and intellect, or is this just a band-aid requirement? That is unacceptable.

Quyen: It is not just about hiring. It is about retention and what you do to retain these valuable assets. It is about making sure that you have an environment that fosters diversity. We also need to think about education and schooling and how we can foster STEM within schools for girls. For some reason, middle school girls tend to drop off around middle school age and think that they are not good at math or science and do not want to follow that path anymore. So, how can we keep them interested? This is something that Swoon Consulting is partnering with one of our board advisors, Melanie Shanks, on. Together, we are working with TechGirlz to introduce some of these concepts and introduce them to math, science, coding, and technology very early on to keep them interested in the subject. Having these types of organizations or working with these organizations is a compelling way to ensure that we have that kind of diversity in thought. When there is a lack of diversity, there is a lack of diverse thought. The danger starts there because we are building technology for the future that is learning, thinking how we think, and only representing the majority. It is not just about girls, but let’s think about underrepresented groups. How do we get ingrained into some of those areas? Another avenue that Swoon is exploring is Girls in Tech, a group that one of our leaders in Toronto, Canada, works with on mentorship programs. So, let’s say that we were able to foster their education, and they are interested in math, science, technology, AI, etc., and then they get into the workforce. Do they have mentors? Do they have advocates who can foster and make sure they grow in their careers? In addition to Black Women in AI, those are two organizations that are important for us to work with. This helps ensure succession planning for girls who are graduating into these fields and their careers.

More than 80% of AI professors are men, only 50% of AI researchers at Meta, and only 10% of researchers at Google are women. So, when you think about professors and how students grow or who they are learning from, they mostly see men. In that case, they will have a specific idea of what this career path will be like. But then, when they join these large companies that are heavily invested in AI, they barely have double digits of women or underrepresented communities. That is such an issue that we have right now. When you ask these companies why they do not hire for more diversity, they say they try, but no one is out there. So, we need to solve some of those areas. I will use an analogy that someone used to describe this to me about fishing. If we are only fishing in places where you already get people from, or you are getting referrals from employees, and you have a majority of white men on your team, you will get more white men. So, we need to think about where we are pulling talent from, where we are investing our time, and how we promote STEM education within these underrepresented groups.

Angle: One of the things I was thinking about the other day is that when I was younger, they would encourage us to take a foreign language, and then it became a requirement to take one because it was so important. One of the things that they forgot was that if I am learning a foreign language, and no one in my family learns the language, who am I speaking the foreign language to? How am I practicing? So, we have to think of that in terms of tech. We have to work in tandem with younger people and the current generation. Our average Black Women in AI members are ages 30 years and older. We cannot just decide that they do not need to have the same skill set that we are teaching the younger generation because then the younger generation becomes the only one who knows the skills. We have to work in tandem to make sure that by the time they get a certain age and enter the workforce, they are not the only ones trying to knock down the doors that should have been knocked down a long time ago. We do not want to repeat that cycle of a younger generation having to fight or break down barriers alone.

Megan: What are some things that organizations should work on or bring attention to, so they are not only able to attract this talent but keep them as well?

Quyen: One of the ways organizations can attract and keep diverse talent is to get involved in the programs where these individuals are. This includes organizations such as TechGirlz, Girls in Tech, Black Women in AI, or finding others that we have not mentioned. Getting involved with these programs will help show your current team members that you genuinely care about it. You do not want to have tokenism. No one wants to be in a company where they are the token Asian, Black person, or woman in what they are doing because you quickly become not just an outcast, but the minority thought. There must be diversity in the room because it represents different thought processes and viewpoints. When it comes to AI, the technology is learning from the algorithms we build, and if we put in flawed data because we are not looking at it holistically, it becomes dangerous from a diverse perspective. Companies must foster diversity and retain this talent by genuinely caring and investing time, money, and everything else into different programs and organizations to help further this cause. Try mentorship programs as well that are specifically for underrepresented individuals within your company that will help foster these individuals and their careers.

Angle: There has to be a shift in the corporate culture. What we did 20 years ago may not work because it is a different world now—for example, internship programs. Interns are not hired based on their work product because they have not worked yet.

Tradition is important, but I believe that if we can find more effective ways to communicate and connect with employees. Sometimes that is just simply being open to more ideas and innovation. When we talk about a shift in the corporate culture, we are talking about a shift in the way that we lead, think, and do business. There are many questions that you should ask yourself at this point. What does that shift look like for your company? What are your employees saying? What are the touch points that you mention that you need to pay attention to? What are your employees saying? Where is your competition? What are they doing differently? When we take a step back and look at it on a larger scale, a shift in the corporate culture is really a call for self-examination. From those in leadership positions and those in the C-suite. When it comes to those that leave your company – how and what are they communicating? It is important that we take a look at their why to see how the organization can improve. Does it align with your mission statement? Does it allow for the sharing of ideas? Does it reflect what you have communicated to the world? Many organizations say that they believe in diversity, equity, and inclusion. However, if those behaviors are not mirrored internally, if the needle has not moved, if there is no actionable plan and someone hired to carry out that plan, then you become disingenuous. Eventually, that will show in the numbers.

Quyen: I had a conversation with a former colleague of mine who is also Vietnamese and an immigrant to the United States like I am. He told me that when he looks at his company, he does not see anyone that looks or thinks like him. It makes him want to leave because he does not see anyone like him in leadership, and it is very demoralizing. Changing the corporate culture has to come from the top down. Your senior leaders have to embrace this change. It is essential to look at your senior leadership team and decide if it looks the way you want it to look and feel. If not, what are the tough changes you need to make?

Angle: Does my leadership team look like the world we serve? You also have to be realistic. That is one of the things that some companies struggle with because they are not practical when it comes to recruiting and hiring diverse talent. Often, they want all PhDs, and that is not realistic because there are a limited number of individuals with PhDs. So, your company will not hire all PhDs from any community. We have to understand what is out there. Not having a Ph.D. does not mean that a person is not capable of successfully doing the job. It is just a requirement that you are asking of someone. Suppose someone came to me and only wanted PhDs because that is all you have in your organization. In that case, that is not realistic because even if I gave everyone in my organization that had a Ph.D., who is to say that they would want to join your organization?

Quyen: I love that you mentioned the Ph.D. comment because, according to Forbes, less than 25% of PhDs were awarded to females and minorities who are historically underrepresented in technology. So, if you were to say that you only want PhDs, that already tells me that you are targeting that 75% that are not women or minorities. So, we have to think about our criteria and the business outcomes we are looking to achieve to make sure it is something you really need and not just a requirement that is already biased.

Megan: What are some of the biases that you have seen in AI?

Angle: One of the biggest things in terms of bias and AI happened with Amazon. Amazon had an algorithm for the recruitment/resume process that weeded out women. It was choosing certain people to recruit, and they ended that situation, but it could happen anywhere if it happened with a major corporation. I do not know who was at the table or behind the scenes, but you would think someone would have caught and questioned it. They may have done that, but I was not there, so I do not know. This could drastically affect people and their livelihood, and we need to be more careful.

Quyen: Another example is your credit score. Many AI tools are driven by data that comes from our society, and our society has built-in biases that we need to consider. When companies are looking to design algorithms for AI, it is not just taking in the data. We need to think about the biases that are already in that data. Angle, in your example, no one probably thought about the data they were putting into their algorithm or how their criteria gave them male respondents. You have to ask, does this help what we are trying to achieve? Is it the business outcome we are looking for? I think some of that came from the fact that women or minorities were not in the room to be able to ask those questions. The same goes for credit worthiness. There is an issue here because we are taking society data, but no one is thinking about what they see in that data that might be represented wrong. We are just taking the data and putting it into an algorithm where machine learning and AI occur. Then, we get credit worthiness for majorities and certain minorities. We need to think about the impacts on society and how dangerous it is for us not to represent the diverse segment of our society. We need to look at how we can safeguard some of these areas and eliminate some biases which can be hard to eliminate. What is essential is that there is diversity in everything, especially when discussing building AI algorithms.

Megan: What are some of the tools companies can implement to reduce or eliminate those biases?

Angle: The best way to mitigate some of the harm being done in terms of bias is to have people in the room who have a voice and authority. I can be in the room, but many of us cannot sway the conversation or have any real input that will make an impact. We have to make sure that the people in the room have the authority to make decisions. In terms of what Quyen mentioned, credit scores can cause economic harm. Credit scores follow us everywhere, whether we are buying a car, home, etc. If my credit score is the lowest it can be, then the wealth gap widens in marginalized communities. We have to be careful with using old data to solve new problems. That is a huge problem when it comes to the criminal justice system and the things they are doing regarding who gets bail. These companies sell technology that determines if you heard shots fired in a specific community but do not know where it came from. Now, you are going in armed, making tensions rise and snowballing into other issues that we do not even want to be a part of. So again, we have to make sure that the right people are in the room and have the authority needed to speak up and actually affect change.

Quyen: A podcast talked about a city government that implemented machine learning and AI for who gets bail to make these decisions less biased. However, it became biased because the numbers were that minorities were not getting bail, and primarily white men were getting it. All because the data going into it was faulty and already biased, to begin with. This algorithm just made decisions, and they let it run until someone said, “Hey, there is something wrong here.” It can be so dangerous, and they just let it decide if someone would get bailed out of prison or not. No tool is out there that is going to say, let’s eliminate bias. If someone creates that, I would probably question it. We have to build safeguards to eliminate and lessen some of our biases. So, what are some of those parameters and safeguards that we have in place right now? We are building some of these algorithms to make sure it is making some of the right decisions and learning from their decisions, so it does not increasingly become more biased. Many companies working in this space have to make sure that we have people in the room who have a voice and influence because unless you can do that, you cannot eliminate bias. No machine or tool will do that for you unless you allow these people and communities to have a voice.

Angle: We have to realize that we are part of that tool. It is not a black box. There is a person behind it. As long as we give the responsibility over to the quote-unquote machine and not the person who is building the machine, we will never get to a less biased place.

Megan: It is essential because some people hide behind the fact that they are using data. It is so insightful to think about what that data holds and the biases within that data.

Angle: One of the most critical questions is, “did you clean the data?” If it is not cleaned out, you know there will definitely be biases there.

Megan: Angle, what was your inspiration for creating Black Women in AI? What are your hopes for the next few years, and where do you see it going?

Angle: I attended an Internet of Things event in Houston back in 2019. When I looked around the room, I did not quite see a complete reflection of myself, nor did I see a reflection of what it would take to provide access and opportunity in this field. I continue to hear it is the fourth industrial revolution, and you cannot have a revolution without Black women. So, with that, we became an organization in August of 2020. We are on four continents – North America, South America, Europe, and Africa. It has been a great 21 months (almost 2 years), and we have some amazing partnerships with Capital One, Nvidia, Expedia, Amazon, Fourth Brain, Dreami, Create Labs, and Swoon’s Black Inclusion Group. It has been just an incredible journey so far.

The ultimate goal is for us to be so successful that we do not need to exist anymore, but in the meantime, we are building a Technology Center in Saginaw, Michigan. We chose Saginaw because it is a marginalized community, and I am from there. I found out that the house I lived in when I was younger was being torn down, and the land was available to purchase. So, I did, and it was personal because my mom’s name was on the deed, and then my name would be on there 20/30 years later. So, we decided to build this innovation center that focuses on our three centers of excellence – education, innovation, and research. We will continue to have our professional development series, work with our partners, and create opportunities for access. We also had the chance to ring the closing bell for the New York Stock Exchange last month, and we are preparing for National Black Women in Artificial Intelligence Month in August.

Quyen: What kind of people are you looking to send through this center?

Angle: We want people first to be an advocate and an ally. We understand the challenges in the world right now, but when people come to the center, we want everyone to have an opportunity, whether they are 8 or 80. We will have those people who were never encouraged in STEM. We will have those who can just create. We want everyone to be able to come in there and learn anything they want to learn, more of a maker space in tech, STEM, and artificial intelligence. We want them to be able to utilize those skills and take them into the world and make the world a better place. So, we will be focusing on those three things mentioned earlier – education, innovation, and research.

Quyen: How can others get involved and connect with you on this?

Angle: They can connect with us on Twitter, @BlackWomenInAI, or email us at contact@blackwomeninai.com. If you are interested in becoming a member or partnering with us, visit our website, https://www.blackwomeninai.com/. We have so many things planned for the future, such as our “In The Company of My Sisters” summer tour from June-August in Houston, Orlando, San Francisco, Washington DC, and New York City.

Megan: Quyen, what are some ways you ensure that diversity is at the forefront of your culture, your consultants, and clients at Swoon Consulting?

Quyen: There are so many ways to think about diversity, and for us, you have to focus on one thing. I love that Angle is focusing on Black Women in Artificial Intelligence. For us, Swoon Consulting and Swoon are woman-owned and women-led. It is imperative to us. When we were building Swoon Consulting and our advisory board, I was very clear in ensuring that our board would be primarily women. I do not want all women because they would not have the diversity in thought that we wish to advise us. We currently have three advisors on our board who are all women. This is fantastic because we know that women leaders, especially top-tier leaders within data and analytics/AI/ML, are challenging to find. I am proud to say that we have Morgan Templar, Melanie Shanks, and Julie Slezak, all powerhouse women in their domain who have been able to do wildly successful things in their careers.

Another thing that I talk about a lot is where we “fish.” When we talk about our RTA team and setting that up, we are looking specifically at Toronto because it is a very diverse city, and we know that there is a lot of diverse talent there. We are also setting up in Mexico. We will be looking into Central America, South America, and Australia in the future. When we say that we want a globally diverse team and to be diverse in thought, our team must represent that. It is at the forefront of who we hire and who is advising us.

Megan: What is your advice to women of all ages who are looking to expand their careers into data or AI?

Angle: Go to YouTube. I know that might not be the popular option, but you have to do your research and figure out where you want to be in this space. You can look at different videos and tutorials to see what you like and what you do not like on YouTube. So often, people will start a boot camp or something like that and spend thousands of dollars to decide that it is not for them. My suggestion is to do your research first and figure out the best path for you.

Quyen:

1. If you see something wrong in this world, be the person who changes or fixes it.
2. Empower yourself because you can do anything you set your mind to.
3. Be a part of the change, and do not wait for someone else to do it.

Meet Our Panel

Read More: