Main content

Equity in AI - stop being condescending and shutting ‘poor’ people out of tech!

Madeline Bennett Profile picture for user Madeline Bennett May 23, 2024
Summary:
How America on Tech and Technovation are training the next generation of diverse tech leaders.

no entry

One of the ongoing concerns around the widespread use of Artificial Intelligence (AI) is bias hardwired into AI programs. As much of the blame for this AI bias has been attributed to a lack of diversity in the teams developing these systems, it’s crucial to attract people from underrepresented groups to the tech sector, especially AI roles.

How best to achieve unbiased AI was the subject of a recent Salesforce event themed around expanding equity and access for the future of technology. This follows last year’s launch of the Salesforce Accelerator – AI for Impact, which gave six non-profits a share of a $2 million fund to develop equitable and trusted generative AI systems to benefit underserved communities.

The equity in AI challenge is one that Technovation has been working on for eight years. The global tech education nonprofit has been offering AI project-based learning to girls since 2016. Tara Chklovski, Technovation Founder & CEO, noted:

Clearly for AI to reach its full potential, it needs to address the bias problem. There are data gaps. Most of the models are created from Silicon Valley or areas in Europe, but the majority of the world lives in Asia and Africa, and we all use these models.

The AI track in the Technovation curriculum, designed to prepare girls and young women from eight to 18 to become tech entrepreneurs and leaders, aims to solve this problem. It’s had thousands of participants from around the world, all focused on building AI tools that tackle local issues. Chklovski added:

For people to see the power [of AI], their data needs to be represented. The Technovation girls are way ahead of this.

Exemplars

Chklovski shared one of her favorite examples: a girl in India was concerned about the water quality in a lake close to her, but didn’t have access to a water quality sensor. She decided to train an AI model to recognize bird sounds, based on her hypothesis that as the lake became cleaner, more bird species would come. As she was training the AI model, she realized that all the bird recognition apps were based on US birds. Chklovski said:

So she built her own data set with Indian birds.

Another example was an 11-year-old girl who wanted to tackle air pollution, a big problem in northern India. She developed an app to help patients determine their risk of lung cancer, training the model based on a tool with US-only patient data. Chklovski added:

Through her app, she's building up a data set with Indian patients’ data. The shortest path to powerful AI models that work for everyone is education.

It’s not just the younger generation that benefit from Technovation’s work. In 2018, it launched the AI Family Challenge, which teaches parents of low-income communities about the technology behind familiar AI tools like Alexa, Siri and facial recognition systems. Chklovski said:

Our mission was to teach these communities how it works, because fear comes from not understanding something. It is critical to have these communities understand what is behind these systems so that you can improve it.

Condescending

The process of getting funding for the program itself also revealed the obstacles for these communities when it comes to inclusion in tech. When Technovation was approaching potential funders, the main response was that those communities are too poor to care about AI. Chklovski said:

Don't be so condescending, give it a shot.

Chklovski gave credit to Google and NVIDIA, which provided funding to launch the AI Family Challenge across around 20 countries, including a refugee camp in Somalia. She added:

The mind-blowing thing across many rural communities across the US, as well as in different countries, was that mothers and grandmothers came to us and said, thank you for teaching us how this works, how the face recognition in our phone works. Nobody ever bothers to think that we are curious, that we care, that we can learn. Too often, education organizations and funders underestimate the power of women and their interest in building technology.

Dual focus

At America on Tech (AOT), an organization set up to prepare the next generation of tech leaders from underestimated communities, there is a dual focus on AI. Firstly, the organization has been adapting the curriculum it’s teaching students to cover the latest AI developments; second, AOT is thinking about how it can leverage AI to improve efficiency in its own internal processes as a nonprofit. That might be for grant writing, reporting or collecting data to highlight its impact more effectively.

It's taken time to progress to this stage, Jessica Santana, AOT Co-Founder & CEO, revealed:

When we think about the introduction of AI at AOT, in the beginning stages we thought, this is scary. Are there robots coming? I imagined the I, Robot movie in my head at one point in time, where my neighbor was going to be an actual physical robot. Now, the uses of AI feel really large in terms of how we can implement the tools and the frameworks within our programs, and then our organization strategy.

AOT is now taking the opportunity to turn its resources to up-skilling communities that have been traditionally left out of the conversation.

The biggest question AOT still has over AI is around equity and access. There could be tremendous repercussions if it isn’t approached correctly, Santana noted:

As we introduce more AI, we also by default create a generation of young people that if they don't have access to those same skill sets, they get left behind.

The organization has also been thinking about the right time to introduce AI to young people. AOT, which works with 16 to 24-year olds, has taken the view that by that age, people have developed critical thinking abilities to be able to understand the issues around the tech. Santana said: 

Maybe if we were working with younger populations of students, we would be a little bit more critical about how we talk about the ethics of AI, and how we introduce integrity-based programming so they’re aware that you don't just use AI for everything, you have to be super intentional about it.

However, these conversations are still very preliminary. AOT has established a small AI advisory board, which has been informing the organization of key developments. But everyone Santana talks to has a different opinion about AI, and there doesn’t seem to be one clear strategy. She added:

Until we know what it is and what it can do, it's hard to put parameters to ensure equity and access is at the forefront, because you can't regulate or put parameters on something that does feel like it's not even structured in and of itself.

While the task of regulating AI is proving complex, one certainty is the need for more women in the industry. Technovation has recently launched the AI Forward Alliance in partnership with UNICEF. The goal is to educate 25 million girls over the next 10 to 15 years so they can train, develop and deploy their own machine learning models and transformative technologies that tackle problems in their communities.

Another aim of the UNICEF partnership is to double the number of women technology professionals from three to six million over the next 10 to 15 years. Programs like Technovation have an important role in meeting this target: according to Chklovski, 76% of the Technovation girls take higher degrees in STEM subjects, and 65% go onto careers in technology. Chklovski noted:

I've been trained as an aerospace engineer and haven't seen too much progress in the number of women in leadership roles in the tech space. The number to me is mind-blowingly low, when we all rely on technology to drive economic progress.

My take

The situation isn’t helped by many of the entrepreneurship programs for women focusing on businesses that are very low impact, low growth, with high competition and no tech focus. As Chklovski noted:

The moment you bring tech in, that's what really accelerates growth. Our ultimate metric of success is increased financial stability and earning power of women. A tech job makes you earn six times more than your non-tech service job.

This realization was also the motivation for Santana launching AOT. Growing up in the Pink House projects in Brooklyn, East New York, Santana’s parents didn't have a lot of resources. She explained:

I got my first line of sight into what it means to be a member of a community while seeing my mom distribute big bowls of rice to our neighbors. Even amid not having a lot growing up, I saw my parents distribute what we did have to others. It created a foundation in my mind that it takes a village to raise our children, we cannot do this work alone to progress our communities forward.

Straight after graduating from Syracuse University, Santana started her career in technology, making about four or five times her family's household income. She added:

I realized this was a pathway towards economic mobility for students that look like me, from communities that I'm from, that oftentimes are not met with the opportunities.

Santana was determined to contribute to ensuring everyone has the same opportunities to succeed as she did, and AOT was the result. The organization started out with 20 students in Brooklyn 10 years ago, and has now expanded to offices in New York, LA and Miami, with over 5,000 graduates of its programs. 

Good exemplars for others to follow on how to tackle the ongoing equity issue. 
 

Loading
A grey colored placeholder image