India introduces Trustworthy AI framework for defence reliability and security

By Staff Report October 18, 2024 2:20 pm

The Evaluating Trustworthy Artificial Intelligence (ETAI) Framework and Guidelines for the Armed Forces were launched in New Delhi. The framework and guidelines are crucial for India’s integration of trustworthy AI into critical defence operations. 

Evaluating Trustworthy Artificial Intelligence (ETAI) Framework and Guidelines for the Armed Forces were launched by Chief of Defence Staff General Anil Chauhan and Secretary, Department of Defence R&D, and Chairman DRDO Dr Samir V Kamat in New Delhi on October 17, 2024. The ETAI Framework and Guidelines mark a pivotal role in the country’s approach to integrating trustworthy AI into critical defence operations.

The Chief of Defence Staff emphasised the necessity of incorporating reliable AI, citing recent global conflicts as a testament to its transformative role in modern warfare. He stressed that it is important to ensure these systems work as intended and are resilient to attacks from adversaries. He congratulated the Scientific Analysis Group for developing the ETAI Framework and praised their efforts in advancing the trustworthiness of AI in defence applications. 

The DRDO Chairman pointed out that reliability and robustness are no longer optional but are essential to prevent mission failures and unintended consequences. He highlighted that AI applications are required to be reliable, robust, transparent, and safe for the success of future operations.

Suma Varughese, a renowned scientist and director general in Micro Electronic Devices, Computational Systems, and Cyber Systems, has introduced the ETAI Framework, a risk-based assessment tool for the defence sector. Its principles, criteria, and measures are equally applicable to other sectors as well. She affirmed that DRDO will play a key role in implementing the framework and assured that all necessary support will be provided to ensure its success.

The ETAI Framework focuses on five broad principles: reliability and robustness; safety and security; transparency; fairness; and privacy. It further defines a comprehensive set of criteria for evaluating trustworthy AI. Complementing the framework, the ETAI Guidelines provide specific measures to be implemented in the AI pipeline to meet these criteria. The framework and guidelines offer developers and evaluators a structured approach to building and assessing trustworthy AI.