Nonprofit Current AI: Pioneering a Free, Universal Web of Intelligence for Everyone

In rural India, a farmer documents a wilting plant but struggles to find information online due to a language barrier. She shouldn’t have to face this challenge.

Current AI, a nonprofit organization, is addressing this issue by creating an accessible public AI framework. Collaborating with Bhashini, the Indian government’s AI language initiative, they launched Suno Sutra—meaning “listening chronicles” in Hindi. This compact device operates offline and supports 22 Indian languages, removing the need for an internet connection. “Countless languages and dialects exist in India, yet AI currently fails to represent them,” stated Current AI CEO Ayah Bdeir in an interview.

Founded in February 2025 by Martin Tisne, Current AI is progressing rapidly. Recently, it allocated $3.2 million in grants to four projects and introduced an open-source AI chatbot during the AI for Good Summit in Geneva.

Bdeir, who previously led AI strategies at Mozilla, joined the organization earlier this year. She is also the founder of littleBits, a company focused on STEM education that reached millions before being acquired by Sphero in 2019.

Operating as a “public-private partnership,” Current AI brings together governments, businesses, and philanthropic organizations to finance technology that serves the public good. The French government initially invested $100 million, joined by contributions from the Ford Foundation, MacArthur Foundation, DeepMind, and Salesforce, totaling $400 million in committed funding. “These aren’t investors; they are funders,” Bdeir clarified.

The organization aims to provide a public alternative to current AI systems, which are predominantly owned by private entities like OpenAI and Google. “If AI truly transforms lives, it should be available to everyone for free, akin to the World Wide Web,” Bdeir asserted.

Many of the world’s languages are at risk of disappearing. “With major AI models being predominantly English-centric, numerous languages, cultures, and communities are left in the shadows,” she said.

When discussing the multilingual strategies of major tech companies, Bdeir highlighted a key difference. “Big Tech focuses on multilingual models primarily for market expansion, often ignoring consent or context,” she stated. The repercussions are tangible: “For Indigenous languages, translations like missionary Bible texts often become training data before community guidelines are established,” she added.

Beyond Language

An AI’s linguistic capacity is merely one facet of its capabilities. “Language transmits knowledge, culture, and identity through generations. If a technology can’t communicate in your language, it cannot preserve your culture,” Bdeir explained.

Bdeir envisions Current AI as an open system inspired by the early days of the web, where advancements benefit all, ensuring no one is excluded and communities maintain control over their own data.

The first grant allocation from Current AI recently involved a distribution of $3.2 million across four organizations in Kenya, Lebanon, and the Brazilian Amazon.

In Kenya, the Masakhane project is focused on creating AI datasets for over 50 African languages in sectors like health and education. Lebanon’s Institute for Worldmaking aims to digitize Arab cultural history into databases that local communities manage. In Brazil, Portal sem Porteiras is creating offline AI solutions with Indigenous Amazonian populations, ensuring data remains within their territory. Additionally, Kenya’s African Internet Rights Alliance is developing tools to hold AI systems accountable across the continent.

Data Ownership Concerns

Regarding data ownership, Bdeir expressed strong views. “Various communities have different models for data ownership, but it’s clear: it shouldn’t reside with a Silicon Valley company aiming to enrich a privileged few,” she remarked.

Current AI’s strategy involves local data storage and engaging community experts before project initiation, implementing consent protocols that allow communities to intervene at any stage.

While no grantees have fully resolved the data ownership dilemma, Bdeir views this as a positive aspect. “They are integrating the question into their work instead of accepting the conventional default where complexity allows tech companies or governments to make decisions for everyone,” she noted.

On the topic of impact from a budget of $3.2 million divided among four entities, Bdeir stated, “Scale isn’t always the goal. That’s the Big Tech mindset. It could be an Indigenous elder in Brazil utilizing a tool developed in Kenya to share ecological knowledge in their native language.”

Developing the Infrastructure

Earlier this month, Current AI unveiled Alpha Chat, an open-source chatbot developed in just seven weeks through collaboration with ten organizations, including Hugging Face, Mozilla, and MIT Media Lab. Each contributor provided elements of the project, such as language models and computational resources.

Additionally, Current AI partnered with Sakana AI, a Tokyo-based startup, to create a shared open-source AI infrastructure that supports the Japanese language and cultures, particularly for communities in the Global South that mainstream AI systems often overlook.