A farmer in rural India photographs a dying plant. She wants to research it, but every major AI tool she could use is built around English. Current AI thinks that’s a problem worth fixing at the infrastructure level, not just the interface level.
The nonprofit, founded in February 2025, is building what it calls open public AI infrastructure: systems that communities can use, adapt, and control without going through a Silicon Valley company. According to TechCrunch, Current AI CEO Ayah Bdeir frames the mission in blunt terms. “If AI is truly a transformative technology, if it’s going to change every aspect of everyone’s life, there has to be a public alternative,” she said. “Like the World Wide Web, available to anyone, for free.”
The French government seeded the organization with $100 million. The Ford Foundation, MacArthur Foundation, DeepMind, and Salesforce have since joined, bringing total committed funding to $400 million. “They’re not investors; they’re funders,” Bdeir said. That distinction matters: Current AI is structured as a public-private partnership, not a company with shareholders to answer to.
The scale of the problem it is trying to address is significant. Every major AI system today, from OpenAI to Google to Anthropic, is owned by a private company. That means the companies decide which languages get supported, which datasets get used, and who profits from the results. Half the world’s spoken languages are facing extinction, and with English driving the largest AI models, most of the world’s languages, and by extension their cultures and communities, are being left out of the AI era entirely.
Bdeir, who joined Current AI in January after leading Mozilla’s AI strategy, draws a sharp line between what Current AI is doing and what Big Tech is doing. “Big tech builds multilingual models to expand their market,” she said, “regardless of consent or context.” The real-world consequences of that approach are concrete. For Indigenous languages, missionary Bible translations often become training data before communities have agreed to anything or set any rules about how their language gets used.
Language, Bdeir argues, is not just a communication tool. “Language is how knowledge, tradition, memory and identity get carried from one generation to the next. So when a technology can’t speak your language, it can’t hold your culture either.” That framing explains why Current AI is not just trying to add more languages to existing AI systems. It is trying to build something structurally different.
One early example is Suno Sutra, Hindi for “listening chronicles.” Built in partnership with Bhashini, the Indian government’s AI language division, it is a pocket-sized offline device that runs AI in 22 Indian languages with no internet connection required. The device is open-source, so developer communities can build on it. It is exactly the kind of tool that could help that farmer in rural India, and it was built without any of the major AI companies involved.
Current AI’s first grant round allocated $3.2 million across four organizations working in very different contexts:
- Masakhane in Kenya is building AI datasets across more than 50 African languages for use in health, farming, and education.
- Lebanon’s Institute for Worldmaking is digitizing Arab cultural history and contemporary practice into machine-readable databases that communities control, not tech companies.
- Brazil’s Portal sem Porteiras is building offline AI tools with Indigenous Amazon communities, keeping data within their territory.
- Kenya’s African Internet Rights Alliance is developing audit tools to hold AI systems accountable across the continent.
None of these projects have fully solved the question of data ownership yet. But Bdeir sees that as the point. “Every one of them has built the question into their work,” she said, “rather than accepting the usual default, where complexity becomes the excuse to let a government or a tech company decide for everyone.” Her position on ownership is direct: “It shouldn’t be a company in Silicon Valley trying to make a select few thousand people wealthier.”
The nonprofit’s approach to data is to store models and data locally, bring in community experts before anything is built, and write consent protocols into the process so communities can stop it at any point. That is a slower and more complicated process than the way most AI systems get built. Bdeir is not apologetic about that.
On the question of whether $3.2 million split across four organizations is enough to matter, Bdeir pushes back on the premise. “Scale is not always the measure. That is the Big Tech paradigm.” Her benchmark for success looks different: “This could look like an Indigenous elder in the Brazilian Amazon using a tool built in Kenya to be able to pass down ecological knowledge in their own language.”
Current AI is also building at the infrastructure level. Earlier this month in Geneva, it launched Alpha Chat, an open-source chatbot put together in seven weeks by a coalition of ten organizations including Hugging Face, Mozilla, and MIT Media Lab. Each contributor brought a piece of the technical stack, covering the language model, safety tooling, and computing power. The point was to show that open AI infrastructure can be built collaboratively, with no single company owning the result.
Current AI has also struck a deal with Sakana AI, a Tokyo-based startup focused on what it calls Sovereign AI. The two organizations plan to build a shared open-source AI stack designed to support the Japanese language and culture, but also communities across the Global South that dominant AI systems have largely ignored. The partnership is a signal that Current AI is trying to build a genuinely global coalition, not just a Western nonprofit with good intentions about the rest of the world.
The organization is moving fast for a nonprofit. Founded in February 2025, it already has $400 million in committed funding, a working hardware product, a launched chatbot, international partnerships, and an active grant program. Whether that pace is enough to matter against the speed and resources of the major AI companies is an open question. But the early internet had scrappy origins too, and Current AI is betting that the same model, open, public, free, can work again.
