IBM’s “Frugal Innovation” Takes Root in Africa


IBM’s “Frugal Innovation” Takes Root in Africa

Combining high and low tech, IBM’s famous R&D lab tackles the challenges of a rapidly urbanizing continent.
Dr. Osamuyimen Stewart, an IBM artificial intelligence researcher from New York, was chosen to head the lab.

IBM Research is one of the world's preeminent R&D labs, having over the decades won five Nobel prizes and invented such core technologies as the disk drive and the relational data base. Last year, IBM announced it was opening a lab in Nairobi, Kenya, its 12th global research facility but its first in Africa. Dr. Osamuyimen Stewart, an IBM artificial intelligence researcher from New York, was chosen to head the lab. "Uyi," as he is known to friends, was born in Nigeria, but has lived in the United States for the last 25 years. He was a panelist at the Stanford Africa Forum, held in March at Stanford GSB, and afterwards, in an interview, described the new IBM lab as more evidence of a vibrant, emerging Africa.

IBM has sales offices all over Africa. Why does it need a research facility there as well?

The way we operate at IBM Research is to look for what we call "grand challenges." We are looking for the hardest problem in the world. That is our DNA. And some of the hardest problems in the world are right there in Africa. So we are excited.

For example?

Transportation. Traffic in Nairobi is a nightmare. As it happens, there are 36 webcams in the central business district, run by a company called Access Kenya that streams them on the web for free. We went to them and asked if we could use their data, and they agreed. The webcams are all very low resolution, but we were able to create an algorithm that detects cars and computes their speed. Then, we used inferences to make predictions for what traffic would be like in other parts of Nairobi, even where there aren't any cameras. And we pass that information along to drivers who subscribe to our Twitter feed on their mobile phones. We built an algorithm that generalizes from a small subset to a much larger set; it's computer science at its best.

Here in Silicon Valley, engineers always want the latest and greatest technology. Yet you seem very excited about what can be done with a technological infrastructure that a lot of people here would consider quite primitive.

We call it "frugal innovation." Frugal innovation asks: In the face of nothing, can I create something? It's a new grand challenge, and as researchers, it gets us excited.

Let me give you another example. Emergency management is almost nonexistent in many parts of Africa. Because there is no 911, when disasters occur there is no one to coordinate a response. And so another thing we've done at our lab is to take advantage of the fact that in Africa there are now more cell phones than adults. Every time there is an incident, people take to Twitter and SMS messaging. It is very noisy data, but we are able to extract useful information from it about what is going on. And we pass it along to policemen and firemen. Previously, during these emergencies, they had been operating in the dark. But now we can use big data to help them.

A lot of people here know about African tech startups involving mobile phones. Are there other kinds of innovation going on as well?

Yes, but it is very unsystematic. You have pockets of innovation, but no one is really thinking how they can translate that innovation and make it commercially successful. That's what's missing. And that's what we are seeking to do with the research lab.

Urbanization is happening at a faster rate in Africa than just about anywhere else in the world. People are leaving rural communities, putting a heavy strain on infrastructure. And as people move out of villages, we are abandoning the mainstay of traditional society in Africa, which is agriculture. Can we begin to devise ways to do "smart" agriculture so that the remaining farmers can provide enough food?

Many countries in Africa have horrible reputations when it comes to corruption. What can a computer scientist do about a problem like that?

Actually, it's a big area of research for us — what the next-generation government sector should look like. We've realized that one of the benefits of E-government delivered via mobile phones is that it would help fight corruption. The more things are digital, the less things are on paper, the fewer opportunities there are for corruption.

The main question being asked at this conference involves who will build the Africa of tomorrow. You work for a big multinational corporation; what role do you think corporations should play in the process?

I like to talk about "co-creation," which is a partnership between multinational corporations and Africans. I don't think that only Africans will build Africa. I believe it will be a cooperative effort of multinational corporations and people on the ground, Africans, all working together.

In Nigeria, for example, 80% of the people are "unbanked." But statistics show that Africans save more than many other people in the world, as a percentage of their income. There are traditions that Africans use to save, but science is unaware of them. Can we use technology to bring them onto the main stage? That is the sort of question that keeps me up at night.

What does IBM opening up a lab in Nairobi say about a new Africa?

I remember 25 years ago, I left because there was no IBM Research Africa. I had to go to Cambridge, and then to the United States. But now we are creating a world-class research environment. So now the best students won't need to go to England or America anymore, because England and America have brought the best lab in the world to Africa.

I am taking my wife, who has never been to Africa, and my kids, who have never been to Africa. We are going because we believe in the change that is happening in Africa and because we want to be part of it.

For media inquiries, visit the Newsroom.
Explore More


The End of Privacy, Part 1 | Art by Tricia Seibold, Photo by iStock/Bim
May 9, 2017

Michal Kosinski: The End of Privacy

In part one of a three part series, a Stanford GSB professor discusses a controversial algorithm that knows more about you than your best friend.


The End of Privacy, Part 2: | Art by Tricia Seibold, hoto by iStock/Bim
May 9, 2017

Michal Kosinski: AI Is Already Smarter Than We Are

In part two of a three part series, a data scientist explores how artificial intelligence developed a mind of its own.


The End of Privacy, Part ​3​ | ​Art by Tricia Seibold, photo by iStock/Bim
May 9, 2017

Michal Kosinski: Can We Out Evolve Artificial Intelligence?

The human mind will outsmart AI in the end, says Michal Kosinski in the final piece of a three part series.