Science and Technology: A Path Toward Gender Equality in Latin America

The Christmas season for Rosa, a Venezuelan migrant and artisan, was the perfect time to sell the handmade bags she made through her business. But it came with a paradox: To produce enough to meet demand, she needed working capital to buy materials and, at the same time, she needed to sell to get the money to buy the inputs.

Not having a credit history, Rosa could only access capital through informal and potentially violent lenders. But when she approached Quipu Bank, everything changed. Using AI algorithms, Quipu Bank evaluated non-traditional data, such as digital wallet transactions and videos of her stock, thus determining that Rosa showed a high payment probability. Today, Rosa has increased her loan capacity from 100 to 1000 dollars.

Just like Rosa's case, Quipu Bank has granted 10,000 loans to small entrepreneurs, exemplifying the potential of AI to close gaps and promote economic inclusion.

The anecdote was shared by Federico Moreno, leader in Artificial Intelligence at said bank, during the recent "Artificial Intelligence: Elements for Inclusive Innovation"talk that took place as part of the GLI 2024 Forum, which also featured Ledenika Méndez, executive director of Digital Inclusion of the Federal Government of Mexico; Paula Cardenau, co-founder of Arbusta; María Eugenia Mereles, senior account executive at Microsoft; and Lance Pierce, CEO of NetHope. The discussion centered on a key question: Could artificial intelligence help us address gender disparities, drive a landscape of equality, and promote inclusion in Latin America?

AI offers enormous opportunities for the social sector. Social impact organizations are already using these technologies to improve their services and reach more people. A notable case is that of Catholic Relief Services, associated with NetHope-a company that works with nonprofits to weave networks that allow them to enhance their use of technologies, and which employs AI to optimize its supply chain and ensure that food reaches those who need it most during food crises. Similarly, the Danish Refugee Council, part of the same network, uses AI to better serve refugee populations, helping them find new opportunities and stabilize their lives on the move.

"These technologies can sometimes be game changers for people who do not have traditional skills training, but can sometimes leave them behind," said Lance Pierce, CEO of NetHope.

Another example mentioned during the talk is Arbusta, a company that uses AI to train young people who lack specific technical training. "We see a huge potential for many people who today do not have access to technological university paths, many young people and especially women in Latin America, to start their careers thanks to having a familiarity with Artificial Intelligence," explained Paula, its co-founder, which is why they have integrated tools such as Mid-Journey to train young people in technical skills quickly, demonstrating how AI can accelerate learning processes and offer new opportunities for professional development.

And for María Eugenia, of Microsoft, that is precisely what this new technology is here to do: "It accelerates us, it empowers us, it gives us the fastest information, but that implies that we have to be mindful to make the right decision and apply it when we deem appropriate."

Because, undoubtedly, the use of this tool also presents significant challenges, especially regarding gender biases. According to figures from the World Economic Forum, 71% of young men use AI, compared to 59% of women. This gap highlights the need to train more women in AI, not only as users but as creators.

Gender Bias in AI

One of the biggest challenges in the use of AI is the elimination of biases, particularly gender biases, that arise from non-representative data or algorithms trained using partial information. An example, widely known but equally impactful, is the difficulty of AI models to recognize facial expressions of Black or other darker-skinned people. Or, for instance, the images associated with terms like "doctor" and "nurse" tend to be biased toward men and women, respectively.

This type of bias perpetuates stereotypes and limits opportunities for women in male-dominated fields. "If we feed it information that carries biases, it will yield biased answers. So I think that it is our responsibility to start feeding and developing processes to operationalize everything related to AI in the right way to reduce or eliminate those biases that we may have," said María Eugenia, senior account executive at Microsoft.

Achievement can only be reached if both the public and private sectors as well as academia are committed to data collection. "Ensuring that we have a dataset that is large and diverse enough is key for statistical representativeness. And there are many paradoxes sometimes hidden in datasets," affirmed Federico Moreno, leader in Artificial Intelligence at Quipu Bank.

So, how can we make AI truly inclusive? It is essential that women also be creators of these technologies. Latin American women, in particular, have a crucial role to play, as their active participation in the creation of AI applications can bring unique perspectives and more diverse data, helping to avoid biases and design solutions that better reflect the needs of the entire population.

It is of the essence that we bring more women into STEM disciplines (Science, Technology, Engineering, and Mathematics) and encourage their involvement in the development of AI. As mentioned by Ledenika Méndez, executive director of Digital Inclusion of the Federal Government of Mexico, "artificial intelligence must be designed by women for women, because, really, who knows us better than we know ourselves?"This perspective is essential for developing AI applications that are meaningful and relevant to Latin American women, who face unique challenges, different from those of women in other regions of the world.