For African AI, Open Source Is Often The Only Option — Not A Virtue

A conversation about who gets to keep their AI proprietary, what that does to African innovation and the role funders can play.

Hanae Baruchel: You’ve said many African innovators don’t choose open source so much as end up there. What do you mean?

Abiodun Adereni: The reason a lot of African tech innovators do open sourcing is because there’s no option. I have to be honest about that. The first time I got into AI was back in 2019, out of work we did with Google — we’d won the Google Impact Challenge in 2018, and they invited us to join the Google AI for Social Goods program. They asked us what opportunity AI could help solve and we built HelpMum Adviser together with Google Research and Vanderbilt University. It helps government institutions properly allocate resources to the people who need them most. Since the time we developed it, we haven’t been able to fully deploy it. The challenge is computing and funding. If an American organization had built that same advisor, I’m sure they’d have it fully deployed by now. The major thing people miss: when you don’t have that option, the only way to get your innovation out into the world at all is to open-source it — make it free for everybody to use.

Baruchel: You’ve used a sports analogy to describe why that doesn’t level the field.

Baruchel: Is there something different about AI than previous technologies when it comes to open source?

Baruchel: So what should be open, and what shouldn’t?

Baruchel: What would it look like if funders got this right?

MORE FOR YOU

Baruchel: There’s a lot of attention right now on AI language capabilities. You say these investments don’t hold as much promise for Africa as they might seem?

Baruchel: How much of the under-investment comes down to funders simply not understanding the problem?

Baruchel: What would you like to see instead?

This conversation is part of an Ashoka series about what works and what’s next in Tech for Human Agency.