Current AI, the nonprofit founded in February 2025 by Martin Tisné, rolled out Suno Sutra at the India AI Summit in February. The handheld device, developed with the Indian government’s Bhashini language division, runs artificial‑intelligence models offline in 22 Indian languages, eliminating the need for an internet connection. The open‑source hardware is intended for developers to customize for local use, Bdeir said.
Offline multilingual device
Farmers and educators in rural areas can now capture a photo of a dying plant or pose a question in Hindi, Tamil or any of the supported dialects and receive AI‑generated guidance without leaving the field. Bdeir emphasized that most major AI systems favor English, leaving hundreds of languages under‑represented. "If AI is truly transformative, there has to be a public alternative," she told TechCrunch.
Funding and grant program
Current AI disclosed a $3.2 million grant allocation to four organizations across Kenya, Lebanon and the Brazilian Amazon. In Kenya, Masakhane will build datasets for more than 50 African languages covering health, agriculture and education. Lebanon’s Institute for Worldmaking will digitize Arab cultural history into community‑controlled databases. Brazil’s Portal sem Porteiras is creating offline AI tools for Indigenous Amazon communities, while the African Internet Rights Alliance will develop audit tools for AI accountability on the continent.
The nonprofit operates as a public‑private partnership. The French government seeded the effort with $100 million; the Ford Foundation, the MacArthur Foundation, DeepMind and Salesforce have pledged additional funds, bringing total committed financing to $400 million. Bdeir, who joined in January after leading Mozilla’s AI strategy and previously founding littleBits, described the backers as “funders, not investors.”
In Geneva last week, Current AI launched AlphaChat, an open‑source chatbot assembled in seven weeks by a coalition that includes Hugging Face, Mozilla and MIT Media Lab. Each contributor supplied a component of the stack—language model, safety tooling or compute power—demonstrating the organization’s modular approach.
Current AI also signed a partnership with Tokyo‑based Sakana AI, known for its “Sovereign AI” work. The two groups will co‑develop an open‑source AI stack designed for Japanese language and culture, with the intention of extending the framework to other underserved languages in the Global South.
According to Bdeir, the ultimate goal is an AI ecosystem modeled on the early World Wide Web: open, interoperable and free for anyone to build upon. By storing models and data locally and embedding community consent protocols, Current AI aims to keep control in the hands of the people who generate the language data. The organization acknowledges that the $3.2 million budget is modest compared with Big Tech spend, but stresses that impact, not scale, defines success.
Cet article a été rédigé avec l'assistance de l'IA.
News Factory APP - actualités agentiques pour booster votre SEO et AEO.