The hype around AI is deafening. Amidst the noise, a crucial question echoes: who benefits? Ai4Dev concerns are we creating a future that the Global South is only an AI consumer or is it going to lead, help define how it’s used, and realize the rewards. More recently, Kava Labs has announced plans to turn its focus towards decentralized AI. This approach offers the upside of transparency and open access, offering a compelling counterpoint to the dark, secretive AI that rules the land today. Yet, can it realistically fulfill its pledge of being inclusive, particularly for developing countries?
Is Decentralization Just Another Buzzword?
Let’s be brutally honest. The crypto sphere is full of these projects that were supposed to move heaven and earth but went on to provide… nothing. Scott Stuart, investor and co-founder of Kava Labs, points to a key shortcoming with the first wave of AI tokens. Their promise was often more exciting than what they delivered. Or are we fated to do so, with “decentralized AI” becoming yet another vacuous marketing catchphrase.
The short answer, per Stuart, is in their usefulness. NFTs, as he dubs them, are just “memes layered on the memes.” Decentralized AI is a powerful tool to tackle the real-world challenges we face today. Imagine farmers in rural Africa using AI-powered tools to optimize crop yields, or healthcare workers accessing vital medical information through AI-driven platforms. These goals aren’t sci-fi dreams! They are very much within reach, as long as the technology is available and adapted to serve local priorities.
Here's the rub: accessibility. Decentralization alone doesn't guarantee inclusivity. Will the Global South really have the infrastructure needed to reap the benefits of all these new advancements? We need reliable internet access, affordable devices, and skilled personnel in place. Or will decentralized AI simply serve to deepen the chasm between the haves and have-nots?
Algorithmic Bias: A New Colonialism?
We need to talk about bias. But AI models trained on data can’t help but reproduce existing societal inequalities if that data does, as the AI will just reflect what’s in it. Think of an AI-powered loan application process that’s been trained mostly on data from majority White countries. Would it continue to do a poor job of credit-rating entrepreneurs in developing countries? Or would it unfairly disadvantage them due to factors outside their control that have nothing to do with their true potential.
This isn't just a hypothetical concern. As a recent Congressional study noted, facial recognition technology is up to 100 times more inaccurate when identifying people of color. Biases are often at the heart of AI systems, which can have devastating consequences. They harden current systems of oppression and create new types of bias.
This is where Kava’s “open-weight” approach makes all the difference. We can detect and fix these biases by preventing AI model parameters from being proprietary. This transparency would be similar to that of Ethereum smart contracts. Open source isn’t simply a nice-to-have; it’s a cornerstone of ensuring that AI is fair and equitable.
Consider this: closed-source AI models are essentially black boxes. Otherwise, we have no way of knowing how they work. For one, we don’t know what data trained them or what biases that data contains. Hashed CEO Simon Kim brings attention to an important point. He cautions that AI may be used in service of corporations or nation-states, sometimes without us even knowing it. This is a pernicious and dangerous course, a course that will seed the ground for a new type of digital colonialism.
Made in the USA, Open to the World?
It is a huge opportunity for the U.S. to really lead in developing AI infrastructure that is decentralized. He imagines this infrastructure as being certifiably “made-in-the-U.S.A.” and powered by open-source funding. This is a compelling vision, but it raises a crucial question: Can a technology be truly decentralized if it's primarily controlled by one nation?
The Global South should be more than just a consumer of American-made AI. It needs to be an active participant in its development, shaping its direction and ensuring that it addresses local needs and priorities. This will demand investment in local talent, infrastructure and social science research. This transformation requires a fundamental change in thinking. Rather than approaching the Global South as an emerging market opportunity, let’s recognize the Global South as our most valued and impactful partner.
There’s an interesting parallel here to the early days of the internet. The West was the original incubator of the internet. It soon became a worldwide movement, putting power in the people’s hands from City Hall to every corner of the earth. Federal support for decentralized AI can help that tremendous potential. It is important that as we do this, we keep it open, accessible, and equitable so that everyone can benefit.
The future of AI is not predetermined. It's up to us to shape it. We can’t make the same mistakes as before, where every new technology has just made things worse on behalf of people who are already disadvantaged. Together, let’s create a future where decentralized, open-source AI tools uplift and empower all of us—no matter where we’re from or who we are. As the largest arena for climate action, the Global South is more than a recipient of aid; it’s an indispensable ally in this journey. Inclusion is not only an equitable imperative, but absolutely necessary for existence. A genuinely global and pro-innovation AI ecosystem hinges on it.
