📰 Key Highlights

Non-profit Current AI is building open public AI infrastructure to solve problems like “Indian rural farmers who want to look up info on dying crops but don’t speak English.” This February, the org partnered with India’s government language division Bhashini at the India AI Summit to launch Suno Sutra (Hindi for “listening chronicles”), an open-source device that runs offline, supports 22 Indian languages, and is open-sourced for the developer community to extend. Current AI was founded by Martin Tisne in February 2025, and CEO Ayah Bdeir, who took the helm this January, previously led Mozilla’s AI strategy and founded edtech company littleBits (sold to Sphero in 2019). The org has been busy lately: last month it disbursed $3.2 million in grants to projects at four institutions, and last week it unveiled an open-source AI chatbot at Geneva’s “AI for Good” summit. Current AI runs on a public-private partnership model funded jointly by governments, corporations, and philanthropic foundations: the French government seeded $100 million, with the Ford Foundation, MacArthur Foundation, DeepMind, and Salesforce jumping in, bringing total commitments to $400 million. Bdeir stresses these backers are “funders, not investors.” She points out that every major AI system today is owned by private companies, while nearly half the world’s languages are on the brink of extinction, and English-dominated large language models push most languages and cultures to the margins. She also criticizes big tech for developing multilingual models purely to expand market reach, often training on data like missionary Bible translations without consent from Indigenous communities. She believes language carries knowledge, tradition, and identity, and Current AI’s vision is to build a public AI alternative — free, open, and accessible to everyone, like the early internet was.

💬 JudyAI Lab Take

Non-profit Current AI is trying to build a public AI infrastructure anyone can use, so language stops being a barrier to information.

Take the Hindi device Suno Sutra — it runs offline, supports 22 Indian languages, and is open-sourced for developers to extend. This highlights an industry blind spot that’s easy to miss: most mainstream AI systems are English-dominated, nearly half the world’s languages are facing extinction, and multilingual model development is often about expanding markets rather than genuinely respecting the knowledge and culture behind those languages. Current AI runs on a public-private model funded by governments and foundations rather than driven by investment returns — in a way, they’re trying to recapture the open, free spirit of the early internet. For AI builders, this is a reminder: accessibility isn’t just a tech problem, it’s a question of who’s included and who’s quietly excluded.

Next time you design a product, take a moment to think about which languages your AI tool supports by default — and who it’s silently leaving out.

📅 Source Info

🔗 Further Reading