Getting your Trinity Audio player ready...
|
The Telecommunication Engineering Centre (TEC), the technical division of the Department of Telecommunications (DOT), has partnered with the Indraprastha Institute of Information Technology, Delhi (IIIT-Delhi), to drive innovation in trustworthy and responsible artificial intelligence (AI) systems.
The two sides signed a memorandum of understanding (MoU), underlining their commitment to pioneering advancements in responsible AI. The agreement focuses on addressing biases in AI systems and promoting fairness in assessments while fostering public trust and confidence in these technologies. The move is in alignment with the government’s strategic goals for the AI industry.
The growing adoption of AI and machine learning (ML) across sectors, coupled with the demonstrated effectiveness of AI/ML in public service delivery and e-governance make it necessary to ensure equality and impartiality in AI systems. Unintended biases in AI applications can lead to a range of ethical, social, and legal concerns.
Bias can infiltrate an AI system in different ways and at different phases of the data and AI lifecycle, potentially impacting other AI components. Algorithms trained on biased data are perceived to contribute to inequalities in outcomes, which puts certain groups at a disadvantage.
Under the partnership, TEC and IIIT-Delhi will create tools to assess bias risks systematically and establish a robust framework to evaluate and certify the fairness and reliability of AI technologies.
Tripti Saxena, Senior Deputy Director General at TEC, said that the collaboration will drive innovative solutions for assessing and rating the fairness of AI systems. The memorandum of understanding acts as a bridge between academia and government entities, encouraging rigorous research and bolstering India’s leadership in the field of AI.
Ranjan Bose, Director of IIIT-Delhi, stressed the transformative nature of the agreement, describing it as a significant stride forward in research, innovation, and technological advancements within the AI field. He informed that the Centre for Artificial Intelligence (CAI) at IIIT-Delhi would lead the execution of the agreement. Academia serves as an impartial partner to TEC in the development and evaluation of robust and fair AI systems, as noted by an official from CAI.
TEC is the officially recognised Standards Setting Organisation (SSO) in the telecommunications and associated ICT sector in India. It is responsible for developing standards for telecommunications and related ICT equipment, networks, systems, and services used in the Indian Telecom Network.
TEC recently unveiled the Standard for Fairness Assessment and Rating of Artificial Intelligence Systems, crafted to instil public confidence in AI. It was developed following extensive stakeholder consultations and the Principles of Responsible AI laid out by the National Institution for Transforming India (NITI Aayog), specifically equality, inclusivity, and non-discrimination.
The main objective of the document is to create a collection of standard operating procedures to assess and arrive at fairness scores for different fairness metrics and a combined fairness score. This standard can be used by governments, businesses, and non-profit organisations to showcase their commitment to promoting fairness.
The standard serves as a tool for validating the risks as part of self-assessments and functions as a form of assurance by a third party when subjected to an independent audit. Globally, independent audits of AI systems are gaining interest, with regulators increasingly mandating compliance and assurance with these audit mechanisms.