India has taken a progressive approach when it comes to the applicability of generative artificial intelligence and data transfer for health and education, but regulation is needed for responsible use and building trust, experts opined in a panel Wednesday.

During the second day of the Global Partnership for Artificial Intelligence (GPAI) Summit in Bharat Mandapam in the heart of Delhi, national and international stakeholders discussed the implications of data usage for generative intelligence to automate tasks, democratize knowledge and make lives simpler.

The panel on Harnessing the power of Data Responsibility, consisting of experts like Sharad Sharma of UN High-level advisory on AI, Professor Jean Gabriel Ganascia of Sorbonne University in France’s capital Paris, and Deputy Secretary of Singapore’s Information and Communication Ministry Hong Yuen Poon, emphasised on the importance of regulation for markets and governments alike in retaining the trust of citizens.

Speaking on the use cases of AI, Sharma said that AI can act as a first level of information for inference.

“It’s the same as a human expert or consultant who can plan your holidays or a doctor who can diagnose, a lender offering loans at a custom interest rate based on your preferences and income. All one needs is a data set of people preferences for deriving patterns,” he added.

Sharma emphasised on the need of multiple AI players, including smaller companies for a techno-legal approach and democratising the technology for public good. He also highlighted the importance of a Digital Public Infrastructure (DPI) model that RBI follows for fintechs, in generative technologies.

Professor Ganascia spoke on the usefulness of conduct for AI, citing the European Union’s Data Act which makes more data available but also regulates through asset preservation and combats bias for responsive data management.

Deputy Secretary Poon shared, “Singapore is a very small country with 5 million (50 lakh) population and a land area of 700 kms, but we have done relatively well in our short history.”

He added that their pragmatic approach for innovation, including a balanced approach for AI that also considers trust with newer approach. He recommended that honesty be incentivised in the governance and industry ecosystem.

Sharma also emphasised on the importance of the Data Protection Act tabled in the Monsoon session this year for Data protection in terms of AI, wherein fundamental right of privacy and protections came to light.

“With reasonable restrictions that are proportionate to innovation, India has worked on data protection principles that are the same (across nations) but India has taken more measures for data as a fundamental right.”

He added that a larger data pool for public infrastructure is better than a smaller data pool in terms of omitting bias and automating decisions, but emphasised on law to be embedded in technology. Open platforms remove entry barriers and harnesses advantages of single platforms, giving smaller players a base to build on, he added.

“We will never be ready for AI, like one is never ready for parenthood, but we must embrace change,” he concluded.

Dr Seydina Moussa Ndiaye, lecturer in Cheikh Hamidou Kane Digital University in Senegal, who guides AI policy for the African union and is working on African AI strategy, spoke on the data problem in the region, lack of resources and capability in Africa and the importance of opportunity to provide data for unbiased models. “With data strategy and AI strategy as our flagship project and work in the segment, it is important that the world provides Africa an opportunity to share data and produce data for building unbiased models,” he said.

The panel also spoke on unlocking the potential of data for health, education or delivery for benefits, and make public data available while it is grounded in principles of human rights.

They also spoke on documented case studies like the UK NHA (National Health Authority)’s partnership with Google DeepMind, the health passbook concept in Taiwan as well as the rapid register for cash transfer in Nigeria and Columbia’s - agri-actionable information.

Dr John Ashley, NVIDIA Chief Architect and Director for AI Technology spoke on the data curation journey, need for proactive data sharing and responsible AI solutions.

“Data curation is a subset of data science and at a practical level, it is important for building systems. Gen AI tools can generate synthetic data to discover gender bias, monitor outcomes, and then adjust the data that was added to blend synthetically.”

Data curation is not one and done, and needs to be sympathetic to privacy and economic agency, he added.

The GPAI is a multi-stakeholder initiative with 29 member countries, which aims to bridge the gap between theory and practice on AI by supporting cutting-edge research and applied activities on AI-related priorities. India is the lead chair of GPAI in 2024, as per the Press Information Bureau.