Originally published as Partner Content by The Financial Times.
When it came time for 17-year-old Esther Egbe – an aspiring engineer living in Texas – to attend college, like millions of students, she needed financial aid to pay for it. Applying for financial aid can be complex and cumbersome, but an AI chatbot called Wyatt, developed by Benefits Data Trust, a US-based non-profit organisation, guided Esther through the process – along with 30,000 other users, more than half of whom are students of colour or from low-income backgrounds.
While much of the discussion around AI today focus on either mitigating harms or reducing barriers to entry for innovators (important considerations in their own right), we haven’t focused enough attention on the end goal: using AI for good.
Safety and accessibility are table stakes, but they’re not enough if economies and societies hope to capitalise on the promise AI offers. We know from experience. The Mastercard Center for Inclusive Growth began its journey to address the growing chasm between those working to solve society’s toughest problems and those with the know-how and assets to unlock data’s potential. At the time, we saw how emerging technologies could be used to build an inclusive economy for everyone, everywhere. It’s why we work to increase access to digital tools and platforms, as well as provide people with the digital literacy skills.
As we celebrate our 10th anniversary in 2024, the Center has been reflecting on how we can help ensure families like Esther’s are able to realise the benefits of the AI revolution. With a decade of lessons at the intersection of technology and social impact, we’ve developed a playbook for how industry leaders can start using AI for good now:
1. Invest in AI tools and research that tackle real-world challenges
Like many data-driven technologies, there is tremendous social good that can come from AI, but it won’t happen by accident. We need to invest in AI tools and research with a specific focus on tackling real-world challenges. For example, as people navigate the transformation AI will bring to the future of work, there’s an opportunity to equip policymakers, employers and workers with tools to upskill and reskill to remain competitive in a rapidly changing digital world. And as the financial sector leverages computer power to improve analytics, they can apply that data to bringing more people into the formal economy and driving inclusive growth.
That’s why we’ve supported FinRegLab in the United States and Women’s World Banking in India, Colombia and Mexico. Both are developing models that responsibly use alternative data and machine learning in credit decision-making for millions of small businesses and consumers. Through our partnership with data.org, Women’s World Banking and the University of Zurich created a “Check Your Bias” scorecard so that financial institutions can assess decision-making about whether and how to lend to women compared to men. This type of research enhances fairness for credit decisions and supports loans for marginalised or vulnerable people, an approach to AI that develops solutions and applies them in order to enable everyone to achieve their full economic potential.
2. Centre data practices around inclusion and diversity
As AI plays a bigger role in business and society, it's critical that AI tools and models be built with representation in mind. Otherwise, AI products will simply replicate biases and historically marginalised groups will continue to be left behind. We’ve already seen this play out: a recruitment AI tool mimicked the decision-making of humans when screening resumes and systemically rejected women applicants and a health care risk-prediction algorithm perpetuated racial bias among Black patients.
We need to address these biases head on, not just because it’s the right thing to do, but because it will make AI technology more useful. A recruitment tool that systematically rejects qualified applicants is simply not effective. That’s why we partnered with Howard University’s Center for Applied Data Science and Analytics, which is working to research algorithmic racial bias and teach future data scientists on how to eliminate bias. Industry leaders should diversify their workforce and data sources, establish internal data governance policies that explicitly centre inclusion and diversity, and regularly audit existing models to identify and rectify biases.
When AI is developed with inclusivity in mind – including with an eye towards training models on inclusive data – AI products can help drive inclusive growth and provide more accurate results for all users.
3. Democratise AI technology
AI has the potential to empower workers and small businesses by improving efficiency, reducing costs, expanding market reach, and fostering innovation and inclusion in the workforce. But we need to be intentional about helping small businesses access AI technology.
Take the example of DataSketch, a Mastercard Strive Innovation Fund grant recipient developing accessible AI tools for Bogota's niche small business markets like design stores, cafes and bookstores. Their tool empowers small businesses and their employees with insights to identify trends in sales, customer behaviour, inventory, and cash flow – and make informed decisions based on real-time data. AI tools like this can go a long way in helping small independent shops – which face increasing competition from larger retailers – stay resilient and grow.
The bottom line is that the AI revolution is here, and it’s up to all of us to determine whether it’s a tool for good. There is no “they” when it comes to developing AI as a force multiplier for positive change and inclusive growth – there is only “we.” From diversifying tech talent and data sources to centring product design around inclusion, we have the power, resources and opportunity to maximise its impact.
Shamina Singh, Founder & President, Mastercard Center for Inclusive Growth and Executive Vice President, Sustainability at Mastercard