The growth of generative artificial intelligence (AI) has attracted interest from key thought leaders in the Western world. However, a recent interview focuses on how innovative AI technology can be used to support the economies of poor countries.

Emad Mostaque, the founder and CEO of the popular AI company Stability AI, shared his opinions about how AI will aid the economies of countries in the Global South in Bloomberg’s latest episode of Exponentially With Azeem Azhar.

AI to Enter the Global South

Mostaque believes that open-source AI models, like his company’s Stable Diffusion and Stable Audio, can easily be trained to adapt to any language and culture.

“My vision is that every person, company, country, culture has their own [AI] models that they themselves build and have the datasets for,” Mostaque said.

Citing Goldman Sachs’ analysis that the implementation of generative AI around the world can help project the global gross domestic product (GDP) to as high as 7% within ten years, Azhar requested the Stability AI CEO’s view on the economic impact of generative AI.

In his response, Mostaque stressed that residents in the Global South, comprising Africa, Latin America, parts of Asia, and others, will “leap forward to intelligence augmentation.”

AI models are trained on billions of datasets existing on the internet using a large sum of money. Mostaque admits that a majority of the datasets used to train existing large models are based on the Western world and their cultures.

Still, he believes that these models can be trained with datasets unique to each country, culture, and language. This can be achieved when the country’s citizens come together to build datasets that harmonize with their country’s needs and use them to train AI models.

Mostaqu stands on the premise that if generative AI is adopted massively by the Global South, it will significantly alleviate the poverty rate and blur the barriers of inequality existing between parts of the world.

The Stability AI founder revealed that his company is working towards using its open-source tools to create AI models unique to diverse audiences. However, he added that these initiatives are still in the research phase and progressing to the engineering stage.