Current AI has partnered with India's government-backed Bhashini to develop a platform that empowers communities to make their own AI devices. The prototype will be showcased at the India AI Impact Summit on February 20th.
Current AI launched at the 2025 Paris AI Action Summit as a first-of-its-kind global public-private partnership that brings together governments, foundations, and the private sector to build public interest AI at scale. This year, Current AI is releasing its first collaborative build and public good, an offline, handheld DIY AI computing device, which will be unveiled at the AI Impact Summit.
The private AI handheld device can describe its surroundings via text or audio across 22 languages spoken in India. The device is entirely self-contained, runs local AI models and requires no connectivity. To enable others to examine, modify and build on it, the device is developed using open-source software and hardware, and its full development instructions will be released in the public domain. The device features a camera, screen, microphone, and speaker, and runs three simultaneous models locally: a Vision model, Bhashini’s Automatic Speech Recognition and Natural Machine Translation model, and a Text-to-Speech model, offering endless capabilities for builders looking to build their own AI devices.
“Our vision is to build a collaborative, collective, and global vision for AI, not one controlled by any single country or company”, said Current AI CEO Ayah Bdeir. “We will do so by partnering with mission-aligned organisations and creators to stitch together the full stack of public interest AI, from hardware to applications, and catalysing a movement for resilient public interest AI.”
AI has a language problem. Most systems are built around English and other Western languages, excluding billions of people worldwide. The Current AI-Bashini prototype takes on linguistic diversity directly, starting in India—where 22 official languages and hundreds of dialects are spoken across 1.4 billion people.
“Bashini has crowdsourced hundreds of thousands of hours of audio data from across the country to build advanced AI models for Indian languages. Having Current AI build an open application using our models and data is a great way to give back to communities, so that they can create their own tools,” added Shri Amitabh Nag, CEO of the Digital India Bhashini Division.
Current AI and Bhashini are also launching a Global Innovation Challenge for communities to imagine what this device could become, explore the boundaries of how open, embodied AI can enable many different futures, shaped by local needs, cultural contexts, and individual imagination, with complete privacy and control.
Ends: For more information, please visit our prototype page or contact media@currentai.org