Mumbai — India is adopting a balanced approach to regulating artificial intelligence, seeking to establish safeguards without discouraging innovation or investment, according to Sandip Patel, managing director of IBM India and South Asia.
Speaking on the sidelines of the IBM Think Mumbai 2026 conference on Tuesday, Patel said the government’s approach reflects an understanding of the need to support technological advancement while addressing the risks associated with AI.
Patel said India is regulating AI without stifling innovation and investment, adding that effective governance will become increasingly important as the technology evolves.
“I think the bigger question that we should be asking ourselves is how do we drive digital sovereignty? And the issue of sovereignty, it’s not just an India issue; it’s global,” Patel said.
Describing digital sovereignty as “Digital Swaraj,” Patel emphasized the importance of maintaining control over AI systems, infrastructure and data while complying with regulatory requirements.
“I think it’s important, both from an AI and infrastructure standpoint. You have to maintain a level of sovereignty that is aligned to the regulations that you are governed by as an enterprise, but also the kinds of workloads that you are, in fact, working with,” he said.
His comments come as the Indian government prepares to release a consultation paper on AI regulation, focusing on safety, workforce development and the prevention of risks such as deepfakes.
Union Electronics and Information Technology Minister Ashwini Vaishnaw has previously emphasized that legislation alone cannot adequately address the challenges posed by AI, calling instead for a combination of technological and legal safeguards.
Patel identified data management, model oversight and transparency as essential elements of responsible AI governance.
He said organizations must ensure that the data used to train and operate AI systems is accurate, properly sourced and used in accordance with applicable requirements.
“Model governance is also the key, which means what models are you using, when are you using it. Are you updating the models? Is your model relevant for what you are trying to use it for? Those are all questions that come into model governance,” Patel said.
He also stressed the importance of explaining AI-generated decisions and preventing unreliable or fabricated outputs, commonly known as hallucinations.
“Do you have transparency in terms of how your models are being used by AI? What data are you using and all the rest of it. But most importantly, the explainability of outcomes. Can you explain your outcomes? Or are you letting your models hallucinate and do whatever they want?” Patel said.
Highlighting IBM’s work in AI governance, Patel pointed to the adoption of the company’s watsonx.governance platform by Indian IT services giant Infosys.
According to Patel, Infosys initially developed its own AI governance system but ultimately decided to adopt IBM’s technology after evaluating its capabilities.
“They had started to build their own governance model. When they saw what we had, they abandoned that, and they adopted our IBM watsonx.gov,” Patel said, adding that Infosys is now using the technology for both internal operations and client services.
Patel said the growing adoption of AI governance platforms demonstrates the increasing importance businesses are placing on accountability, transparency and regulatory compliance as they expand their use of artificial intelligence. (Source: IANS)
