Current AI, a nonprofit founded in February 2025, is pursuing open public-interest AI infrastructure through partnerships, grants and open-source tools. The organization, led by Chief Executive Ayah Bdeir, says its aim is to create a public alternative to AI systems run by private companies, with communities retaining control over their data.
At the India AI Summit in February, Current AI partnered with Bhashini, the Indian government’s AI language division, on Suno Sutra, an offline, pocket-sized AI device designed to operate in 22 Indian languages. The device is open-sourced for developer communities to build on. Current AI also recently introduced an open-source chatbot at the AI for Good Summit in Geneva.
The French government provided an initial $100 million for the initiative, according to Bdeir. Current AI says commitments from the Ford Foundation, MacArthur Foundation, DeepMind and Salesforce have brought its total committed funding to $400 million; Bdeir described the contributors as funders rather than investors. Its first grant cohort allocated $3.2 million among four organizations working in Kenya, Lebanon and the Brazilian Amazon.
The supported projects include Masakhane’s work on datasets in more than 50 African languages for health, farming and education; an Institute for Worldmaking project to make Arab cultural history and contemporary practice machine-readable under community control; offline AI tools with Indigenous Amazon communities through Portal sem Porteiras; and AI accountability audit tools being developed by the African Internet Rights Alliance. Bdeir has argued that multilingual AI built by large technology companies may not sufficiently account for community consent or context. Separately, analyst Nathan Lambert has warned that possible future policy action could restrict advanced open-weight models, though he noted that reported White House discussions had no official confirmation.
