New Delhi: IndiaAI has selected five transformative projects under its ‘Safe & Trusted AI’ initiative, focusing on deepfake detection, bias mitigation and AI penetration testing. These projects aim to make AI in India safer, transparent and more trustworthy.
The selected projects emerged from the second round of Expression of Interest (EoI) launched on 10 December 2024, which received over 400 proposals from academic institutions, start-ups, research organisations and civil society. A multi-stakeholder technical committee rigorously evaluated submissions, identifying five initiatives that combine resilience testing, bias audits and robust AI evaluation tools to support responsible AI development and deployment.
Under the Deepfake Detection theme, three projects were chosen: Saakshya, a multi-agent, RAG-enhanced framework by IIT Jodhpur and IIT Madras; AI Vishleshak, developed by IIT Mandi in collaboration with the Directorate of Forensic Services, Himachal Pradesh, which strengthens audio-visual deepfake and signature forgery detection; and a real-time voice deepfake detection system by IIT Kharagpur.
In Bias Mitigation, Digital Futures Lab and Karya will evaluate gender bias in agricultural AI models, creating Digital Public Goods for benchmarking and fair data practices. For Penetration Testing & Evaluation, Globals ITES Pvt Ltd and IIIT Dharwad will develop Anvil, a tool for testing and evaluating large language models and generative AI systems.
These initiatives underscore IndiaAI’s commitment to building a resilient AI ecosystem. By combining academic expertise, industry innovation and civil society participation, the mission seeks to ensure that AI systems deployed in India are secure, reliable and inclusive, aligning with the nation’s vision of ethical and responsible AI.
Also Read –
Govt to Set Up Over 500 Data Labs Nationwide Under IndiaAI Mission
