The Ministry of Electronics and Information Technology (MeitY) is launching a programme to encourage institutions like IITs and also startups to build artificial intelligence tools and framework to mitigate bias and ensure algorithmic fairness.
The project will be taken up under the aegis of the IndiaAI programme and the ministry intends to provide grants to institutions which can build 10 such tools based on the theme of “Responsible AI”.
“As AI integration grows, India aims to invest in agile mechanisms for indigenous tools and assessment frameworks, contextualized to its socio-economic realities. Under the pillar of ‘Responsible AI’ of National Program on Artificial Intelligence (NPAI), IndiaAI – IBD, under Digital India Corporation, launches an initiative for research project-based funding,” says the expression of interest floated for the project.
The ministry is looking for solutions in machine unlearning, which plays a crucial role in rectifying inaccuracies, biases and outdated information; synthetic data generation tools, which create fabricated data instances that mimic the characteristics of genuine data; algorithm fairness tools, which ensure automated systems are treating all individuals fairly and so on.
The ministry said institutions such as IITs, National Institute of Technologies, Indian Institutes of Science Education and Research, other central/state government universities, research and development organisations, private universities and colleges are can participate in the programme and receive funding.
“The Academic/R&D Organizations should have the pre-existing Lab infrastructure (Lab having Workstations, Servers, Project Staff etc.) for project implementation and for training of students under the programme, along with relevant research publication,” the EoI said.
Interested institutions and startups will have to submit proposals, which will be evaluated by a committee of experts.
The other themes that institutions and startups can take up are developing AI ethical certifications, which are formal assessment and recognition processed that validate that AI systems, services or organisations adhere to established ethical principles and guidelines.
Institutions can work towards developing AI governance testing frameworks and algorithmic auditing tools. While the former will ensure compliance with governance policies and ethical guidelines, the latter is a process of evaluating the impact and behaviour of machine learning models.