How a data strategist is fighting gender bias in Generative AI

Generative AI has taken the world by storm, and chatbots like Google’s Bard and OpenAI’s ChatGPT are out in force. Asha Saxena is already a step ahead, with a new bestseller book, The AI Factor, and through her work as founder and CEO of Women Leaders in Data & AI (WLDA), a private network for senior executive women tech leaders. She’s also an adjunct professor at Columbia University. Asha shares her thoughts on potential applications of AI,  how businesses can leverage AI for innovation and growth, and her mission to remove the gender bias in data.

Archie: 2023 is certainly shaping up to be an exciting year for AI. Did you expect this?

Asha: In the past when we talked about AI it was just about research and development. Nobody got to see what was really behind the closed doors there. Come November 2022, I think this was such a remarkable time when ChatGPT came out, and in the first five days it had a million users. Now that you’ve put AI into the hands of the people, they can now engage with Generative AI technology, and they’re fascinated with it. You’re engaging with human AI and having these conversations, and businesses are asking, ‘How can I create an API (application programming interface) through ChatGPT and be able to improve our technology?’ But, there’s still a large gap between where we’re with AI now to where we’re hoping to go.

Archie: How is Generative AI connected to the continuing evolution of Big Data and Machine Learning? 

Asha: The fundamental of AI is really Big Data and the variety of data—the text, images, audio and video. We connect Generative AI to Big Data because Big Data stays essential to any kind of Machine Learning model. Generative AI is leveraging this vast amount of data and is getting trained on it. The quality of the output produced by Generative AI depends on the quality of the input. So the training data set dictates the output that you’re going to get. The machine learns from Big Data and then it gives you the output to give you the prediction or that action that’s like human intelligence: a recognized pattern to categorize things. 

Archie: Could you have been interviewed today by a chatbot instead of me? 

Asha: AI is not necessarily putting you out of a job, but your job might become different and look different. You’ll be using human intelligence to combine with machine intelligence to do more creative work. Now we’re in a world where we don’t have to write code anymore. You have machines that can generate algorithms. But yes, the large use cases revolve around content generation. Anything creative like screenplay for movies, music composition, artwork for designs, and personalized content for targeted advertising. 

Archie:  What excites you the most about possibilities from Generative AI?

Asha: - What I’m personally excited about is healthcare and education. We’ll start seeing a lot more on the personalization of the treatment plans for patients. This entails analyzing personal medical records, being able to prescribe a treatment, and then especially in how we engage with a patient around care. There’s a humorous example of a robot reminding an elderly woman patient to do her physical rehabilitation tasks. The robot says, ‘Come on, let’s go for physical therapy’ and the patient rolls her eyes as in, ‘It’s just a robot—what does it know?’ and the robot says, ‘Oh, you’re rolling your eyes at me…you don’t believe me because I’m a robot’ and the woman says, ‘Oh my god, you’re annoying.’ 

With education, I look at the possibilities of personalized education plans, as in, ‘What are you interested in?’ This approach can replace traditional learning tracks. Generative AI can now personalize your course material based on what you’re actually interested in. How cool would that be?

Archie: A solution to all of those undecided majors?

Asha: Exactly!

Archie: Your new book The AI Factor helps businesses leverage AI to drive success. What are the typical roadblocks facing companies when it comes to commercializing AI?

Asha: For a lot of organizations, the AI experiment is just sitting in Stage 1 phase, as in an experiment or R&D component. It’s comprised mainly of beta testing, and organizations are not really moving to the next level in terms of building a path to commercialization. Organizations become so fascinated with the technology that they start playing and experimenting with it. There’s a lack of understanding of what exactly AI can do. When you cannot make the business strategy that aligns to the technology you’re adopting, then there’s a breakdown and you’ve created a problem with your whole project. A lot of organizations are also very siloed in their departments and that leads to further misalignment in overall business strategy. Then there’s the usual organizational culture change concerns, and the worries about the project becoming super expensive. 

Archie: You’ve created a business model for AI that you’ve termed the Data Power Canvas. What are the principles behind this model?

Asha:  The reason I call it Data Power Canvas is I wanted the canvas to find the power that business data provides. In my diagram, there are four columns. The first two columns are really about data, as in, ‘Do I have the right data? What data do I have in hand? Do I trust that data?’ And then, ‘Can that data tell me what problem I’m solving?’ For example, ‘Do I have the market’ or ‘Do I want to expand in the market for what I’m delivering?’ When you have the data that can drive you to those questions, then you deliver the AI Factor, where the data can give you the performance to impact the output either for your product, your customer, or your price and cost. 

Archie: Your book also talks about ethical and sustainable AI. What do we need to worry about as Generative AI takes a prominent place in society?

Asha: What worries me is the bias, because we live in a world that’s biased. The Internet is biased, the data is biased, and so the output that we get for Generative AI will be biased, even with ChatGPT. There are gender-based assignments that affect this data, from male and female responses. Responsible AI is really important. There needs to be boundaries as to how people ethically use data as applied to AI. The common person’s underlying fear of this technology revolves around privacy and overreach. That’s what we as professionals in the space think about and worry about, even before Generative AI became an everyday discussion.  

Archie: You created an organization and private network, Women Leaders in Data and AI, or WLDA (pronounced Wilda) to address head-on the inclusion and gender inequity bias. How are you doing that?

Asha: WLDA focuses on the inequity around data that affects generative AI. We need to be more cautious about our differences. If we don’t create a world that’s more inclusive and more diverse, we won’t have the data and output that we’re looking for. I think that’s the bias that’s getting caught in these algorithms. 

I’ve spent 30 years in technology and had never thought about parity and equity before. During the pandemic it seemed everything was going digital. I thought, ‘Oh my god, we’re creating this digital world that’s designed by men, but only 18% of women are writing these algorithms.’ And that was worrisome, because if we don’t have diverse leaders designing these programs, we’re going to have problems. We’re writing history today, and it’s so important that we have women at the table creating the digital world for the next generation. My goal is to bring together senior leaders of Fortune 1000s with the mission to create a digital world that’s a better place for having parity and equity. We’re creating a world with more women at the C-suite and senior levels in technology roles around, digital, data and AI. It’s a network community of WLDA Leaders, WLDA Nation for more junior professionals, and a mix of mentorship models to address these diversities around gender, ethnicity, and color at all levels. 

Archie:  Through your Entrepreneur-in-Residence stint at Columbia Business School, are you seeing a change in the types of careers that business school graduates are pursuing as a result of new opportunities in data?

Asha: Columbia Business School was known for preparing its students for investment banking and consulting. That was the main popular area that graduates were going into, but most recently it has become big on innovation and entrepreneurship. It’s big on AI and Big Data and disruptive technology, as they are actually teaching business school students technology topics. 

Archie: In a world of fast-changing tech, what does the future hold?

Asha: Augmented Intelligence fascinates me, where humans and AI are partners together to create efficiencies. It’s going to be more collaborative now. Right now, you have this atmosphere where technology is sitting in the corner, and you have solutions that come out but now it’s really going to be more engaging. Augmented Intelligence is about helping humans, not replacing humans.