CHENNAI: The Centre for Responsible Artificial Intelligence (CeRAI) has started to create a regulatory framework for AI models for easier and wider adoption of the technology.
The CeRAI, a first-of-its-kind research centre in the country on AI ethics at Indian Institute of Technology, Madras campus, is aimed at carrying out fundamental research and advocacy for developing responsible AI-based solutions. It will define bias in Large Language Models (LLMs), creating test mechanisms for identifying bias, creating tools for training unbiased models with a biased dataset, fixing liability in case of a problem in an AI-based system. The centre will also focus on creating policy guidelines for the governments and chart out a path for enforcing and monitoring those regulations.
The regulation will be in healthcare, education and manufacturing sectors and the centre later plans to enter into law and other areas. The idea is to focus on the area where the cost of mis-prediction is very high, explained Prof V Kamakoti, Director of IIT-M. “It is a process to develop an AI quality framework for our country, as it can vary for every geography. We will undertake research in developing framework, training and spreading awareness in this topic for various stakeholders, including artificial intelligence tool developers, institutions running them and end users. It will have an impact on the future of this technology.”
He stressed that regulation will fall into three layers - Generic regulation by the Union government, sector-specific regulation from industry bodies and voluntary regulation of the organisation using it. Prof Kamakoti suggested that job loss fears related to AI are unwarranted as they have many undigitised areas in the economy.
Government should regulate AI-based tools and solutions and not development in the area, said B Ravindran, head of CeRAI, IIT-M. Industries can self-regulate to some extent but it won’t be enough, he said, adding the regulation should not hinder the growth of the industry.
Google has provided a million dollar grant to CeRAI, and it was formally inaugurated last month.While the excitement around the AI-based models has surged Chat GPT and Google’s Bard, critics are raising concerns over it and urging the governments to regulate the industry.