Interview | Hannah Finch, Editor, Leesman Insights
The insatiable need of deep-learning algorithms to train on massive quantities of labelled data has spawned an entire industry of human labellers. Leila Janah, CEO of Samasource, talks about how the rise of AI is creating jobs for the world’s poorest people.
Samasource is the rarest of organisations in Silicon Valley: it’s structured as a hybrid organisation, which means it’s a for-profit that is majority owned by a non-profit. Like many successful tech companies, it was born from an innovative, disruptive idea with the potential to change the world. What sets it apart is that it does not exist just to become a unicorn and have a huge IPO, but to transform the lives of the world’s poorest people by giving them jobs in AI.
The story of Samasource is intricately linked to the story of its founder, Leila Janah. When Janah was 17, she taught English in Ghana as part of a student-volunteer programme. Many of her students were blind and living without running water or electricity. But despite overwhelming challenges and hardships, they managed to make it to school every day eager to work and learn.
Janah went on to study African Development Studies at Harvard, and on graduating in 2005 landed a job as a management consultant. One of her first assignments was to manage a call centre in Mumbai. She noticed that the majority of call centre workers were from well-educated, middle-class families.
“In the last decade, Samasource has created over 10,000 jobs in data services and AI for some of the poorest people in Kenya and Uganda.”
While outsourcing was providing millions of jobs in India, it wasn’t helping the country’s poorest people. Having worked with people from extreme deprivation in Africa, Janah was convinced that they were more than capable of thriving in a call-centre type of setting. In 2008, at 25, Janah turned that idea into Samasource, a non-profit that connects poor people to digital jobs.
In the last decade, Samasource has created over 10,000 jobs in data services and AI for some of the poorest people in Kenya and Uganda, the majority of whom are women. Each of those 10,000 people support an average of four income dependents, so more than 50,000 people have directly moved out of poverty as a result of Samasource.
We talked to Janah about how she’s made it work, and what her plans are for the future:
Q: Samasource is such an incredible model, but not everyone was convinced in the beginning. How did you get started?
We were challenged by everyone in the beginning. People said: ‘Even if you manage to get this off the ground, you’ll never be able to scale the business.’ The best way to face those kinds of doubts is to prove people wrong with data. As we started getting our first contracts, I personally reviewed every piece of data services work we did and checked the quality. We grew organically from there, and now solve training data challenges for 25% of the Fortune 50.
When a customer was unsure about working with us, we would do a free trial for them, saying ‘if you’re happy with the result then keep using us’. That approach really works. Once people saw the high quality of the work and the proof of the social impact there was a ‘wow’ factor that set us apart. From there we were able to grow quickly – we have offices all over the world and count Google and Microsoft as clients.
Q: And what an impact you’ve made in the last decade! Tell us about how the Samasource model works.
In the last ten years internet access across Sub-Saharan Africa has made it possible for entirely new kinds of work that didn’t exist before. For example, there’s a massive need in the algorithms of the future to use images to train computers to do what humans used to do.
Those computer vision processes rely on human annotation and all kinds of input data. To understand our company, you need to understand a bit about how machine learning works: it starts with input data, this can be video, images, snippets of texts, and then we need to precisely label that data. Machines learn the same way humans learn when we teach a toddler the difference between a tree and a shrub. That information goes into algorithm development and then one of the end users of Samasource will push that out into a product that reaches the customer.
The head of AI at Tesla has said that ‘human labellers are the programmers of the future’. And if there’s one company that understands the appetite for this kind of data, it would be Tesla.
“Most people think that AI means that humans will be totally automated out of the value chain; that’s not what we’re hearing from the biggest tech companies in Silicon Valley.”
Most people think that AI means that humans will be totally automated out of the value chain; that’s not what we’re hearing from the biggest tech companies in Silicon Valley, where our headquarters is based.
Q: Samasource is certainly proving that AI has vast career potentials within it. Tell us about your AI training centres in Nairobi and North Uganda.
Since we started a decade ago, we’ve become the largest data services company in East Africa, employing about 10,000 people, all of them from extremely low-income backgrounds, most of them earning about $2 a day before Samasource.
Northern Uganda is mainly known for the horrific civil war that was waged by the Lord’s Resistance Army, led by Joseph Kony. Most people wouldn’t imagine that in war-torn Gulu you could set up a tech centre. Our Gulu centre is off-grid, has solar panels on the roof, and has employed over 500 people inside refurbished shipping containers.
There are so many benefits to this approach to giving work: it’s much lighterweight than a factory. I don’t have to worry about getting goods through customs, about importing raw materials. It’s a much more efficient way to create large numbers of jobs than many forms of traditional commerce. That’s the promise of a digital economy.
“It’s a much more efficient way to create large numbers of jobs than many forms of traditional commerce. That’s the promise of a digital economy.”
Q: And what do your workers’ lives look like before and after Samasource?
On average, our workers are starting at about $2/day, which means that they are victimised by all kinds of entirely avoidable suffering solely due to poverty. One of the biggest risk factors for disease and violence in the developing world is living in informal housing, in a slum. We find that people are not eating nutritious food – they are eating sugar cane as their primary source of calories.
For the 10,000 workers we’ve employed, we’ve increased their income by 500%. This means that they can get health insurance, access to education, proper sanitation. This is the power of work over charity.
Q: One of the impressive things about Samasource is that you’ve put equal consideration into the working environment at all of your locations – from San Francisco to North Uganda. Why is that important to you?
Samasource has an impact working model, and our impact is in the job-creation aspect of what we do. I passionately believe in creating living-wage jobs and hiring people from backgrounds who wouldn’t traditionally be able to get those jobs.
Hiring people from really economically depressed backgrounds means there are not only minorities but there is also a gender component: more that 50% of our workplace is female. We look at all aspects of impact sourcing – considering not just poverty but other forms of marginalisation.
When your employees come from challenging backgrounds, the workspace that you’re putting them into is really important. Many of our employees live in the slums, which are incredibly difficult environments to survive in. For our employees who live in the Kibera slum in Nairobi, it means there’s an open sewer outside their home. There’s no running water. There’s no proper sanitation.
Our impact model is not just about a paycheck, it’s about improving lives on a daily basis. The office environment is one of the best tools we have to do that. That’s why we strive to make Samasource workplaces inspiring. It’s a bright and airy space with encouraging words written on the wall. We offer nutritious food, evening transportation, health benefits and many other benefits. We also incorporate sustainable building principles and employee health and wellness practices into our workspace design globally.
“Our impact model is not just about a paycheck, it’s about improving lives on a daily basis. The office environment is one of the best tools we have to do that.”
Q: One of the natural benefits of working in a well-resourced workplace is the natural community that forms around it. How important is it that your employees feel connected to each other?
I definitely try to foster an atmosphere where people feel friendly with their colleagues, but I don’t want to force people into social engagement. My primary goal is that people will get the benefit out of their work and feel like they are growing and achieving things and meeting personal life goals.
The best social interactions are ones that spring up naturally. Our team in Kenya often play soccer behind our building. Because they initiated it, they own it, which is so much better.
We do sponsor the ‘SamaCup’ every year with both our Kenyan and Ugandan teams. There are all kinds of games and dancing, it’s great fun. And we do a global management retreat with our team in the Bay area which includes attendees from our offices worldwide. Last year, that included two of our folks from Uganda who started as agents then became project managers and now they run projects of over 50 people. It was their first time leaving Africa and taking a plane; they were a brilliant addition.
Q: It’s fantastic that your employees have so much scope for job progression. Are there any employee stories that stand out for you from the last decade?
A lot of people assume that the reason people are in poverty is because they are uneducated and therefore not qualified for well-paid jobs. This often is not the case – in Kenya unemployment rates for youth are around 70%.
Let me tell you about Martha. She was abandoned as a child and ended up at an orphanage. When she was 18, she had to leave and had no training and nowhere to go and ended up in sex work. One of the nuns from her orphanage heard about Samasource and knew that Martha was an incredibly competent young woman, so they put her in touch with us and she got hired. She did really well and worked for us for two and a half years before joining the marketing team for a travel company in Kenya.
I saw her last year and she is a completely transformed person to the girl who joined Samasource. Her demeanour is like night and day compared to how she was before, she is full of confidence and hope now. She’s caught the entrepreneurial spirit and is working on creating her own fashion line. Confidence is one of the biggest gifts we can give our employees. Especially for women who come from very low-income backgrounds like these slums, they’ve been told so many times that they’re not capable or not worthy, and so to give them the confidence to build their own futures is so exciting.
Q: What’s next for Samasource?
We are planning to expand more in Sub-Saharan Africa. There’s a lot more demand for our work and we’re working with some of the biggest tech companies now – Microsoft, Google and the major automotive companies.
The next question for us is: ‘how do we scale up?’