TLDR: India has announced five innovative projects under the Safe & Trusted AI pillar of its IndiaAI Mission, aimed at strengthening a secure and responsible artificial intelligence ecosystem. These initiatives will address critical challenges such as deepfake detection, gender bias in AI systems, and generative AI security, following a rigorous selection process from over 400 proposals.
In a significant move towards fostering a secure and responsible artificial intelligence landscape, India’s Ministry of Electronics and Information Technology (MeitY) has unveiled five cutting-edge projects. These initiatives, selected under the Safe & Trusted AI pillar of the ambitious IndiaAI Mission, are designed to tackle pressing challenges within the AI domain, including the proliferation of deepfakes, inherent biases in AI systems, and the security vulnerabilities of generative AI.
The announcement, made on October 8, 2025, marks a crucial step in India’s commitment to building a robust and ethical AI framework. The projects were chosen from a pool of over 400 proposals submitted during the second round of the Expression of Interest (EoI), which was launched in December 2024. A multi-stakeholder committee, comprising leading technical experts, meticulously evaluated submissions from top academic institutions, innovative startups, and established research organizations before shortlisting the final five.
Each of the selected projects brings a unique focus to the overarching goal of safe and trusted AI:
1. Saakshya: A collaborative effort by IIT Jodhpur and IIT Madras, this project introduces a multi-agent, RAG-enhanced framework specifically designed for advanced deepfake detection and governance.
2. AI Vishleshak: This joint initiative between IIT Mandi and the Directorate of Forensic Services, Himachal Pradesh, aims to significantly improve the detection of audio-visual deepfakes and handwritten signature forgeries. Its focus includes adversarial robustness, explainability, and domain generalization.
3. Real-Time Voice Deepfake Detection System: Led by IIT Kharagpur, this solution is engineered to combat the growing threat posed by synthetic voice manipulation in real-time scenarios.
4. Evaluating Gender Bias in Agricultural LLMs: A partnership between Digital Futures Lab and Karya, this project is dedicated to creating Digital Public Goods (DPG) for benchmarking and promoting fair data practices, particularly in the context of large language models used in agriculture.
5. Anvil: Developed by Globals ITES Pvt Ltd in collaboration with IIIT Dharwad, Anvil is a specialized penetration testing and evaluation tool tailored for generative AI systems, ensuring their security and reliability.
Also Read:
- Indian Competition Watchdog Highlights AI Risks, Advocates for Big Tech-Led Industry Oversight
- India’s Artificial Intelligence Market: A Decade of Exponential Growth Predicted by Industry Reports
The IndiaAI Mission, operating as an independent business division under MeitY, serves as the primary implementation agency for these initiatives. Its core objectives include democratizing access to AI technologies, cultivating technological self-reliance within the nation, and championing ethical and transparent AI practices across all sectors of society. The Safe & Trusted AI pillar, under which these projects fall, is instrumental in ensuring the responsible development, deployment, and adoption of AI by fostering indigenous tools, guidelines, and governance frameworks. Furthermore, the IndiaAI Safety Institute has been established under this pillar to advance science-based and indigenous research on AI governance, pioneering innovative approaches to AI safety relevant to the developing world.
