Christopher Marquis

Skip to main content

Sama Validates Its Ethical AI Supply Chain Through A Randomized Control Study With MIT

Sama Validates Its Ethical AI Supply Chain Through A Randomized Control Study With MIT

by ESG Business Institute -
Number of replies: 0


Artificial intelligence is an emerging force in the business world that has the potential to either replace humans in certain industries or empower humans with better tools, depending on how the technology is utilized. Sama, a training data and validation company based in San Francisco, believes AI can enhance how we work and is advocating for “human-in-the-loop,” a work model that requires human involvement even with advanced technology.

Since inception, Sama has been on a mission to provide opportunities to individuals in underserved regions across the world. Through its impact sourcing model, Sama has helped more than 55,000 thousand people lift themselves out of poverty. The company provides living wages that are about 4.3x higher than those typically received in the regions where they operate. 

While there are many companies that do such good in the world, it is harder to rigorously verify that the company is actually having a positive effect on the individuals and communities it serves. To determine the effectiveness of its impact sourcing model, Sama conducted a three-year Randomized Control Trial (RCT) of its model in collaboration with the Massachusetts Institute of Technology (MIT) and Innovations for Poverty Action. 

For an RCT, it is essential to have a well identified control group with which to compare the “treatment” – in this case receiving training and working for Sama. For Sama’s study, individuals from low-income and marginalized backgrounds in Nairobi, Kenya were split into three groups, the Control Group received neither training nor employment at Sama, Group 1 only received training, and Group 2 received training and worked at Sama. The outcomes studied included employment rates, average earnings, and individual well-being. 

After the three year study, the results clearly showed that individuals that received both training and work at Sama (Group 2) had lower unemployment rates and higher average monthly earnings as opposed to the Control Group and Group 1. Furthermore, the impact of training and employment at Sama was even more significant for women than men.  

I recently spoke with Sama CEO Wendy Gonzalez as part of my research on purpose-driven business and to learn more about the RCT and how Sama is using its advanced technology to enhance human-centered work and provide job opportunities to marginalized communities. Sama is also a B Corp, having met rigorous social and environmental standards as verified by the nonprofit B Lab. 

Christopher Marquis: Can you please describe Sama’s model and how the company is set-up to have social and economic impact on its employees and communities in which it operates?

Wendy Gonzalez: Our late founder, Leila Janah, was passionate about building an impact-led technology company. As a result, Sama has been guided by a mission to provide unprecedented opportunities to individuals and communities in underserved regions worldwide since its inception.

Towards these goals, Sama has positively impacted more than 55,000 people by opening doors to long-term careers in AI and data training.  

We’ve achieved this by employing and investing in the careers of professionals across the globe, including countries historically left out of the digital economy, such as Uganda, Kenya and Costa Rica. Through initiatives like ‘SamaU’ (Sama University), we’ve remained committed to investing in our employees long-term, enabling them to acquire high-paying skills that will open doors for them to upskill both within and outside of Sama.   

As a female-founded and led company, we’ve maintained a focus on creating opportunities for women within the tech industry. Today, over 50% of all company agents in Africa are women who have been able to support themselves and their families with the skills and employment they have received at Sama. 

In addition to ensuring our employees have key skills that allow them to grow professionally, we also provide living wages that are on average 4.3X higher than typical wages offered in these areas. 

Marquis: You recently publicized results from a 3 year study of your model. Can you describe why an RCT was necessary to show the results of your model and the design of the RCT? 

Gonzalez: We knew we were positively impacting people’s lives through our own data collection, however, a few years ago, we had reached a point where we wanted to have it officially validated. Toward this goal, we decided to pursue a Randomized Controlled Trial (RCT) study, widely considered the gold standard in research. Our goal was to validate the results of Sama’s impact-led business model on improving people’s lives globally. 

To achieve this, we partnered with The Massachusetts Institute of Technology and Innovations for Poverty Action to study three groups of individuals in Nairobi, Kenya. One group was a Control Group that did not receive training or the opportunity for employment at Sama. Group 1 only received training at Sama but did not receive an opportunity for employment, and Group 2 underwent training and subsequently was given the opportunity to work for Sama. We then spent three years surveying individuals from within these groups to accurately assess factors such as employment status, earnings, and the industry they worked in after leaving Sama. 

Marquis: What are the results of the study, and can you please provide some mini-case studies of employees and your communities that illustrate them? 

Gonzalez: The most important finding from our RCT study was the validation of our impact model. The study found that individuals who received training and the opportunity for employment at Sama reported higher employment rates and earnings than both of the other groups studied. Specifically, we found that people who were trained and had an opportunity to work at Sama (Group 2) reported wages that are 40% higher compared to the control group and unemployment rates were 10% lower. 

The study also found that Sama’s impact model was even more effective for women. The data shows that women in Group 2 reported average wages 60% higher than women in the control group. Furthermore, women in Group 2 exhibited a 270% increase in monthly earnings after joining Sama. 

I want to emphasize that these outcomes are more than numbers on a page. As a result of our mission, people in our operating regions are able to better support themselves and their families while gaining skills and opportunities that are unprecedented in these communities. 

To demonstrate Sama’s impact on its employees at an individual level, I want to share the story of one person who has benefited immensely from joining our team. Before joining Sama, Habiba was working several casual jobs to support her family of eight including customer care roles, data entry jobs at a cyber cafe, and freelance work. She heard about our company and free training program, ‘SamaU’, and decided to enroll. For two weeks, Habiba learned fundamental skills like how to use a computer, how to send emails, and best practices for interviewing. Immediately upon finishing the program, Habiba was brought on full-time and is now one of the pioneers in GPL. As a result, Habiba can now live more comfortably and described her experience with Sama by saying, “At the end of the month, I get to send something to my Mum and the kids to help out with the bills. My previous workplaces had a lot of pressure but at Sama, it’s more relaxed and comfortable to do your work.” This is just one example of how our impact mission is helping people lift themselves out of poverty every day by giving them the training and opportunities they need to succeed. 

Marquis: How does Sama's impact model differ from other data training or AI companies (ex. Amazon's Mechanical Turk) who have been known to power their business models through "ghost workers?" 

Gonzalez: Other AI companies have been in the news recently for unethical treatment of employees and exploitation of “ghost workers” in underserved regions. Hearing these people’s stories is extremely disheartening. Findings like these demonstrate how crucial impact models like Sama’s are so that employees and their families are not caught in cycles of exploitation. 

Sama is different from others in the space given that our impact model has been the foundation of our business since inception, and our goal has always been to empower individuals and their communities. With the completion of our RCT study, not only have we demonstrated our commitment to impact, but we have also proven that our training and employment programs are truly bettering outcomes for people globally. Now that the study is complete, we are proud to share the data and analysis to clearly demonstrate the ways in which Sama’s mission is making a real difference around the world with our proven ethical AI supply chain. 

The last time we spoke, Sama had just become one of very few AI companies to receive a B Corporation Certification. Unlike our competitors, Sama has taken it upon itself to receive third-party validations like RCT and B Corp that hold our company accountable to keep impact and ethical practices central to our business model and values. The AI industry has a problem and it is up to companies like Sama to recognize the issues at hand and proactively work to change them. 

Marquis: This RCT study shows that women are especially impacted by Sama's model. Why is that and why is it central to your mission? 

Gonzalez: Yes! As previously shared, our RCT study showed women who received training and the opportunity for employment at Sama experienced a 270% increase in monthly earnings and exhibited average wages 60% higher than women in the control group. 

Achieving these goals of empowering women has been incredibly important to us. Oftentimes, women in our operating regions are disproportionately disadvantaged in comparison to men. Recognizing this issue, Leila set out to provide women and their families with better opportunities, meaningful work and living wages.  

Receiving results from our RCT study has been an amazing experience because I know Leila would be proud to know her mission to “give work” is empowering thousands of women globally and the data to back this up is better than we could have ever imagined.