Current AI Launches Open Tools to Democratize Multilingual AI Access

Current AI is developing open-source tools and datasets to extend AI capabilities into languages and communities that major commercial systems have overlooked. The nonprofit, established in February 2025, has already produced an offline device capable of running in 22 Indian languages and an open chatbot assembled from contributions by multiple organizations.

Funding and Early Milestones

The French government provided the initial $100 million, with additional commitments from the Ford Foundation, MacArthur Foundation, DeepMind and Salesforce bringing the total to $400 million. Ayah Bdeir, who became chief executive in January after heading AI strategy at Mozilla and previously founding the education company littleBits, has overseen rapid grant activity. Last month the organization distributed $3.2 million across four recipient groups working in Kenya, Lebanon and the Brazilian Amazon.

Those grants support language datasets for health and farming in more than 50 African languages, digitization of Arab cultural records under community control, offline tools for Indigenous groups in the Amazon, and audit mechanisms for AI accountability across Africa. The projects emphasize local data storage and consent protocols that allow communities to stop work at any stage.

Recent Product Releases and Partnerships

In February the group collaborated with India’s Bhashini language division at the India AI Summit to create Suno Sutra, a pocket-sized device that operates without internet connections. The hardware is fully open-sourced for further development by local programmers. Last week Current AI released Alpha Chat, an open-source chatbot built in seven weeks by a coalition that includes Hugging Face, Mozilla and the MIT Media Lab, and announced plans with Tokyo-based Sakana AI to create a shared stack supporting Japanese and other underrepresented languages.

The organization’s stated goal is to produce public AI resources comparable to the early World Wide Web, where improvements remain accessible and communities retain oversight of their own information rather than ceding control to private companies. Work continues on consent frameworks and localized models, with the next steps focused on expanding the coalition and testing additional language and cultural datasets.