Nonprofit AI Seeks to Create a Free Global Network for Artificial Intelligence
Image Credits:Getty Images
Bridging Language Gaps in AI: The Vision of Current AI
A farmer in rural India captures an image of a wilting plant, eager to research its condition online. Unfortunately, she encounters a barrier as she doesn’t speak English. This situation highlights the pressing issue that many face in an increasingly digital world: access to information should not be limited by language.
Addressing Language Accessibility
This is precisely the challenge that Current AI, a nonprofit organization, aims to tackle by developing open and public AI infrastructures. In February at the India AI Summit, Current AI partnered with Bhashini, the Indian government’s AI language division. Together, they introduced Suno Sutra—an offline, pocket-sized device that runs AI in 22 Indian languages without needing an internet connection. “In India, there are hundreds of different languages and dialects, and currently AI does not adequately represent them,” explains Ayah Bdeir, CEO of Current AI, in a TechCrunch interview. The device is open-sourced, inviting developer communities to contribute and enhance further.
Rapid Progress in AI Development
Founded in February 2025 by Martin Tisne, Current AI’s mission is evolving swiftly. The organization recently allocated $3.2 million in grants to four diverse projects focusing on language representation and cultural preservation. Highlights include the launch of an open-source AI chatbot at the AI for Good Summit in Geneva, underscoring their commitment to bridging technology and community needs.
Ayah Bdeir, who joined the team in January after a successful tenure leading Mozilla’s AI strategy, comes equipped with entrepreneurial experience from founding littleBits, a STEM education company that reached millions of children before its acquisition in 2019.
Current AI embodies a “public-private partnership,” uniting governments, corporations, and philanthropic organizations in the realm of public-interest technology. With initial backing from the French government for $100 million and further support from the Ford Foundation, MacArthur Foundation, DeepMind, and Salesforce, they are approaching a remarkable total funding commitment of $400 million. “They’re not investors; they’re funders,” Bdeir clarifies.
The Need for Public Alternatives
The fundamental challenge Current AI confronts is clear: the majority of AI systems, including those from OpenAI and Google, are held by private companies. Bdeir asserts, “If AI is genuinely a transformative technology, then there must be a public alternative available to everyone, free of charge, akin to the World Wide Web.”
As half of the world’s spoken languages face extinction, the stakes are high. Bdeir notes, “With English dominating the largest language models and AI systems, many other languages—and the cultures and communities they represent—are being overlooked.”
The Pitfalls of Big Tech Multilingualism
When discussing Big Tech’s efforts toward multilingual AI capabilities, Bdeir differentiates between their motivations and those of Current AI. “Big Tech often builds multilingual models primarily to expand their market, often without the necessary consent or context,” she states, warning that Indigenous languages may be used as training data prematurely, without communities establishing rules.
More Than Just Language
The ability of an AI to converse in various languages is merely part of the equation. Bdeir emphasizes language’s role in transmitting knowledge, traditions, memory, and identity through generations. “When a technology cannot communicate in your language, it cannot encapsulate your culture,” she asserts.
Current AI envisions an open system akin to the early web, where collective improvements benefit all users without exclusion, enabling communities to retain control over their data.
Community-Focused Grant Programs
In its first grant round, Current AI distributed $3.2 million to four organizations across Kenya, Lebanon, and the Brazilian Amazon. Each initiative is designed to empower local communities:
- Masakhane, Kenya: Focuses on creating AI datasets across 50 African languages for healthcare, farming, and educational purposes.
- Institute for Worldmaking, Lebanon: Works to digitize Arab cultural history and contemporary practices into machine-readable formats that communities control, not tech companies.
- Portal sem Porteiras, Brazil: Collaborates with Indigenous Amazon communities to develop offline AI tools, ensuring data remains within their territories.
- African Internet Rights Alliance, Kenya: Focuses on developing auditing tools to hold AI systems accountable across the continent.
The Question of Data Ownership
Addressing data ownership, Bdeir emphasizes, “Various models exist on data ownership within communities, but it shouldn’t be companies in Silicon Valley profiting at the expense of others.” Current AI’s approach is community-driven; they store data and models locally, consulting with community experts before implementation. Consent protocols are integrated into the development process, ensuring communities can withdraw at any time.
While no grantee has entirely resolved the issue, Bdeir views the focus on this question as a significant step forward. “Each organization is engaging with these complexities rather than accepting the typical default,” she states, challenging the notion that confusion should allow tech companies or governments to dictate solutions.
Impact Beyond Scale
When evaluating the effectiveness of a $3.2 million budget across four initiatives, Bdeir articulates a broader perspective: “Scale is not always the measure. This could be an Indigenous elder in the Brazilian Amazon using a tool created in Kenya to transmit ecological knowledge in their language,” she says, illustrating the meaningful connections and knowledge-sharing possible through their initiatives.
Collaborative Innovation in AI
Earlier this month, Current AI unveiled Alpha Chat, an open-source chatbot developed in seven weeks through a collaboration involving ten organizations, including Hugging Face, Mozilla, and MIT Media Lab. They each contributed components of the technological stack, such as language models and safety measures.
Furthermore, Current AI has partnered with Tokyo-based Sakana AI, known for its Sovereign AI efforts. This collaboration aims to build a shared open-source AI stack that not only supports Japanese languages and culture but also extends its reach to communities in the Global South that mainstream AI platforms often overlook.
Conclusion
Current AI is at the forefront of cultivating an inclusive and equitable AI landscape that respects linguistic diversity and cultural heritage. By prioritizing community engagement and open-source practices, they are not only addressing the challenges of language accessibility but also empowering marginalized communities to maintain control of their identities and data. Through their initiatives, Current AI envisions a future where technological advancements enrich rather than exclude, paving the way for a truly participatory digital era.
Thanks for reading. Please let us know your thoughts and ideas in the comment section down below.
Source link
#Nonprofit #Current #racing #build #World #Wide #Web #free
About The Author
