A new non-profit organisation, Current AI, is spearheading an ambitious project to construct an open, public artificial intelligence infrastructure, aiming to create a 'World Wide Web of AI' that is freely available to everyone. Founded in February 2025 by Martin Tisne, the organisation is moving at pace, driven by the belief that a transformative technology like AI requires a public alternative to the existing proprietary systems.
Current AI’s mission extends beyond mere technological development; it seeks to address the significant linguistic and cultural biases present in current AI models. CEO Ayah Bdeir, who joined in January after leading Mozilla’s AI strategy, highlighted that with English dominating large language models, a vast number of the world's languages and cultures are being left behind. The organisation views language as central to knowledge, tradition, memory, and identity, asserting that an AI unable to speak a local language cannot truly represent a culture.
Remarkable progress has already been made. In February, at the India AI Summit, Current AI collaborated with Bhashini, the Indian government’s AI language division, to launch Suno Sutra. This pocket-sized, offline device operates AI in 22 Indian languages without needing an internet connection, demonstrating a commitment to local accessibility. Furthermore, last month, Current AI allocated £2.5 million in grants to projects across four organisations in Kenya, Lebanon, and the Brazilian Amazon, focusing on building AI datasets for local languages, digitising cultural history, and developing offline AI tools for Indigenous communities.
The initiative operates as a 'public-private partnership,' drawing funding from a diverse range of sources. The French government provided initial seed funding of £79 million, with additional contributions from the Ford Foundation, MacArthur Foundation, DeepMind, and Salesforce, bringing the total committed funding to £315 million. These entities are described as funders rather than investors, underscoring the non-profit's public interest mandate.
Current AI's vision is an open system mirroring the early internet, where technological advancements benefit everyone, access is universal, and communities maintain control over their own data. This approach directly contrasts with the commercial models of major tech companies, which Current AI argues often build multilingual models primarily to expand market reach, sometimes without adequate consideration for consent or context, particularly concerning Indigenous languages.