A farmer takes photos of dying plants in rural India. She wants to look it up on the Internet, but she doesn’t speak English. she doesn’t need to do that.
This is the kind of problem a nonprofit called Current AI is trying to solve by building an open public AI infrastructure. At the India AI Summit in February, it partnered with Vashni, the Indian government’s AI language arm. The result is Suno Sutra, which means “listening chronicle” in Hindi, a pocket-sized offline device that runs AI in 22 Indian languages and doesn’t require the internet. “India has hundreds of different languages and dialects, and AI doesn’t represent them right now,” current AI CEO Aya Budair said in an interview with TechCrunch. This device is open source and available for construction by the developer community.
Founded in February 2025 by Martin Tisne, this non-profit organization is moving quickly. Last month, the company allocated $3.2 million in grants to projects across four organizations. Just recently (last week) we announced an open source AI chatbot at the AI for Good Summit in Geneva.
Bdeir joined Mozilla in January after leading Mozilla’s AI strategy. She previously founded the STEM education company littleBits, which reached millions of children before selling it to Sphero in 2019.
She told TechCrunch that today’s AI functions as a “public-private partnership” that brings together governments, businesses, and philanthropy to fund technology in the public interest. The French government seeded Current AI with $100 million, with participation from the Ford Foundation, MacArthur Foundation, DeepMind, and Salesforce, bringing total committed funding to $400 million. “They’re not investors. They’re funders,” Budair said.
The problem we are trying to solve is straightforward. From OpenAI to Google to Anthropic, all of today’s major AI systems belong to private companies. “If AI is truly a transformative technology that will change every aspect of everyone’s life, then there has to be a public alternative to it,” Buder said. “Like the World Wide Web, it’s free and available to everyone.”
Half of the world’s spoken languages are at risk of extinction. “And with English driving the largest language models and AI systems, the majority of the world’s languages, and as a result, cultures and communities, are being left behind,” Buder said.
When asked about major tech companies’ multilingualization efforts, Bouder made a clear distinction. “Big tech companies build multilingual models without consent or context in order to expand their markets,” she said. The results are tangible. “For indigenous languages, the translation of the Bible by the missionaries becomes the training data before the community establishes any rules,” she said.
Not just language
The ability for AI to speak language is only part of what it needs to learn. “Language is how knowledge, tradition, memory and identity are passed on from one generation to the next. So if technology can’t speak language, it can’t preserve culture,” she said.
Her vision for modern AI is an open system modeled after the early web. There, improvements benefit everyone, no one is locked out, and the community remains in control of its own data.
Current’s first cohort grant round, announced last month, included deploying $3.2 million to four organizations in Kenya, Lebanon, and the Brazilian Amazon.
The project in Masahane, Kenya, involves building an AI dataset across more than 50 African languages for health, agriculture, and education purposes. Lebanon’s Institute for World Creation is digitizing Arab cultural history and contemporary practices into machine-readable databases managed by communities (rather than technology companies). Portal sem Porteiras in Brazil works with Amazonian indigenous communities to build offline AI tools and store data within their territory. The African Internet Rights Alliance in Kenya is also developing auditing tools to hold AI systems accountable across the continent.
Who owns the data?
On the issue of data ownership, Budair didn’t mince words. “There are different models and proposals for who owns the data in different communities, but one thing is for sure: it shouldn’t be a Silicon Valley company trying to enrich a select few,” she told TechCrunch.
The nonprofit’s approach is to store models and data locally, bring in community experts before building anything, or write consent protocols into the pipeline so the community can stop the process at any time.
None of Current AI’s grantees have yet to fully solve this problem. But Budere thinks it’s important. “They’re all building questions into their work, rather than accepting the usual default of using complexity as an excuse to leave everyone’s decisions to governments and tech companies,” she said.
“Size isn’t necessarily the measure, and that’s the Big Tech paradigm,” Budair said of how much progress could be made with a $3.2 million budget split between four organizations. “This could look like indigenous elders in the Brazilian Amazon being able to use tools built in Kenya to convey ecological knowledge in their own language.”
Building the stack
Earlier this month in Geneva, Current launched Alpha Chat, an open-source chatbot assembled in seven weeks by a coalition of 10 organizations including Hugging Face, Mozilla, and MIT Media Lab. Each contributor brings part of the stack, including language models, safety tools, and computing power.
Current AI also signed an agreement with Sakana AI, a Tokyo-based startup known for developing something called Sovereign AI. The two organizations will build a shared open-source AI stack designed to support not only the Japanese language and culture, but also communities across the Global South that have been largely ignored by dominant AI systems.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
