Regulating Open Data_ Principles Challenges and Opportunities

and Mr. Arun Prabhu, Partner and Co -Head, Digital and TMT Practice, Cyril Amachan Mangaldas. We also have the distinguished presence of Ms. Asha Jadeja Motwani, Founder of the Motwani Jadeja Foundation. So, before we begin, I intend to invite Dr. Shashi Taru to deliver a keynote address, but given the extraordinary significance of the discussion today we are having, I quickly created a scenario where this Global AI Summit is expected to be attended by many individuals, and many have attended. I have created a scenario where I transport you back to the Prime Minister’s office in the United Kingdom. Imagine Jim Hacker, PM of UK, Sir Appleby, the Cabinet Secretary of UK, as well as Sir Bernard Hooley, the Principal Private Secretary, are here to attend this.

So, I am going to create a scenario for three minutes. Bear with me. hacker, the Prime Minister says, Humphrey, I’ve decided this is the most important session at the entire Global AI Summit. And Humphrey says, Prime Minister, with respect, there are panels on frontier AI, sovereign computing and semiconductor strategy. Hacker, exactly. All terribly glamorous, but this one is about open data, the plumb. Without it, the rest is just PowerPoint. Bernard, yes, Prime Minister, they are discussing whether India should move from voluntary open data initiatives to a statutory regulatory framework that actually requires government bodies to share standardized aggregated data sets. Sir Humphrey says, requires? Hacker, yes, Humphrey, safeguards, incentives, accountability, coordination between ministries, even defined economic models for access, free, paid and restricted tiers.

Sir Humphrey replies, Prime Minister, the beauty of open data policies is that they are aspirational. Once you introduce statutory mandates, you risk consistency. Hacker, That’s the point. They’re arguing that without a legal backbone, participation is uneven. Some departments share, others don’t. No uniform standards, no enforcement. Investors get nervous. All developers complain about unreliable data sets. Bernard, and apparently, high -quality public data improves evidence -based policymaking, targeted welfare delivery, and even capital formation. Humphrey replies, yes, Bernard, but also improves scrutiny. Hacker, Humphrey, they’re not just talking about dumping spreadsheets online. They’re debating architecture, secure environments, anonymization protocols, synthetic data, interoperable standards. And Bernard replies, and ensuring privacy and copyright protections don’t clash with open data objectives.

Sir Humphrey says, Prime Minister, when privacy, innovation, geopolitics, and economic growth are all mentioned in the same regulatory framework, one usually convenes a task force to study it indefinitely. Hacker replies, but that’s precisely what they are avoiding. They are asking the real question, should there be regulatory teeth so that government data sharing isn’t optional goodwill but institutional obligation? Hacker replies, they are also discussing geopolitical standards and safeguards, access restrictions. In other words, minister structured openness rather than chaotic transparency. And Humphrey replies, structured openness is merely closeness with better branding. Hacker, Humphrey, if AI is the future, the data is raw material. And if government holds the richest data sets, then refusing to regulate sharing properly is like building a digital economy and locking the warehouse.

And Hacker replies, Humphrey, that’s why this is the most important panel. Everyone is discussing what AI can do. They are discussing what governments can do. I want to stop here and invite my dear friend and mentor, Dr. Shashi Tharoor, to deliver the keynote address. Thank you.

Thank you. That was delightful, Raj. I was terrified for a minute that you were going to get me to play Sir Humphrey or something. But this is a pleasure to join you all this evening at the India AI Impact Summit 2026 and to share a few reflections on the subject that Raj has so cleverly animated for all of you, exploring a regulatory framework for open data. When artificial intelligence is no longer a distant frontier of innovation, it is rapidly becoming the operating system of our modern society. What was once theoretical is now embedded in our markets, our governance systems, and increasingly our personal choices. A nation’s digital footprint, a sort of triad of three Cs, commerce, communication, and cognition, is now its primary source of wealth.

We’re often told, almost as an article of fiction, that data is the new oil. Yet, as Chris Miller reminds us in his compelling account in Chip War, the real constraint of the AI age is not the volume of data, but the power to process it. That single sentence punctures a convenient myth. It is, excuse me, it tells us that abundance alone does not confer agency, and that openness without capacity can entrench inequality as easily as it can enable progress. The decisive question, therefore, is not how much data exists, but who controls its use, who extracts its value, and who is left behind. Seen in this light, the regulation of open data is not a technical footnote.

It is a question of power, shaping sovereignty and surveillance, innovation and inclusion, freedom and fairness in our digital age. age. It’s a privilege to share this platform with such a distinguished and accomplished group of colleagues under the stewardship of my good friend Rajkumar, whose intellectual leadership has shaped conversations on law and global governance. I’m honoured, of course, to engage alongside Shival Shroth, Asha Jadeja Motwani, Arun Prabhu, Mirama Vedushree, Aireena Ghosh, and my parliamentary colleague, though in a different house, Sasmit Patra, individuals whose expertise across law, technology, policy, industry, and democratic institutions has profoundly shaped the very debates we’re having today. To speak in the company of such authority is both an honour and a responsibility, and so how, let me ask, might we craft a regulatory framework for open data that is equal to the ambitions we all have and the anxieties many of us are expressing about the AIA?

So to begin with, we must be clear about what we mean by open data. At its most basic, it refers to data that is made accessible for use, reuse and redistribution with minimal legal or technical barriers. Yet in the context of the AI age, open data is far more than a question of access. It’s a statement of intent about how knowledge is shared, how power is distributed and how societies choose to govern the informational foundations of innovation. When designed thoughtfully, open data becomes more than a technical tool, it becomes public infrastructure. It strengthens transparency in government, levels the playing field in markets and creates genuine avenues for citizen participation. But when released without clarity, safeguards or purpose, as Bernard pointed out in Raj’s presentation, it risks becoming little more than symbolic.

A sort of symbolic nod to open data. It can turn into an unguarded channel through which value, agency and even sovereign control quietly drift elsewhere. We all know that open data can be genuinely transformative. We’ve seen how making government data publicly accessible can strengthen democratic accountability, whether it’s citizens tracking public spending, researchers analysing welfare delivery or civil society organisations flagging gaps in implementation. India’s own open government data platform has been used to track welfare coverage and expose leakages in implementation that might have otherwise remained invisible. But the value of open data extends beyond transparency alone. When the United States chose to release meteorological data freely, they did more than increase transparency. They laid the groundwork for entire private ecosystems in weather forecasting, logistics, insurgency and security.

They laid the groundwork for entire private ecosystems in weather forecasting, logistics, insurance and risk assessment. What began as public infrastructure became the foundation for commercial and technological growth. Its importance becomes even clearer in times of crisis. During the COVID pandemic, openly shared health data and public dashboards enabled faster responses, improved coordination across agencies, and supported more informed public debate. So if we take these examples, the lesson is consistent. When data is treated as shared infrastructure rather than as a guarded asset, it lowers barriers, improves decision -making, and enables societies, particularly in the developed world, I’m sorry, in our developing world, rather, to turn information into durable capacity. And yet, my dear friends, openness alone is not a panacea.

Open data, poorly structured, can generate new vulnerabilities, even as it promises transparency. without safeguards openness may devolve into tokenism data sets released without context quality control or enforceable standards or worse into asymmetrical extraction there is a trilemma of digital governance digital ascendancy digital capitulation and digital sovereignty today most of the world’s large cloud servers and advanced artificial intelligence systems are owned and operated by a small number of technology companies based primarily in the United States and parts of Europe this is digital ascendancy this means that data generated in developing countries whether it’s mobility data from ride sharing apps digital payment transactions agricultural statistics or health records is often stored, processed and analyzed on infrastructure located abroad When that data is then used to train AI systems, improve algorithms or develop commercial digital services the profits, patents and technological advantages tend to accumulate where the platforms are headquartered not where the data is originally generated Put simply, the location where data is produced is not necessarily the location where value is created This is where the question of data sovereignty arises If countries do not invest in their own digital infrastructure and regulatory capacity the benefits of open data can accrue disproportionately outside their jurisdiction One -sided concessions on digital taxation and digital trade are a form of data capitulation Indonesia and Malaysia have succumbed in their trade agreements with the US We must not Thank you This dynamic is increasingly playing out in real policy debate It is visible in digital trade negotiations where restrictions on data localization or limits on source code disclosure can narrow the policy space of developing economies seeking to nurture domestic digital industries.

It is also evident in the market concentration of hyperscale cloud providers whose global dominance shapes where data is stored, processed, and ultimately valorized. The issue is not cross -border data flows per se. Digital cooperation depends on them. The concern is whether openness is reciprocal and capacity enhancing or whether it systematically positions some countries as suppliers of raw data while others capture downstream gains in artificial intelligence, advanced analytics, and platform governance. An instructive example, when the U.S. sought to compel the divestiture of TikTok, TikTok was the first to be introduced into the market. TikTok was the first to be introduced into the market. Its demands included mandatory data localization, majority U.S. ownership in the restructured entity, and U.S. control over source codes.

This is data sovereignty on steroids, and it’s exactly what the rest of us seem to be only able to aspire to. The answer, therefore, is not to retreat from openness, but to shape it deliberately. If openness without strategy creates imbalance, then openness with guardrails can create resilience. A credible regulatory framework for open data must begin with clarity of purpose. Why is this data being released? For whom and under what safeguards? It must ensure strong anonymization and privacy protections so that transparency does not come at the cost of individual rights. Closely linked to this is the principle of consent and control. Individuals and communities should have meaningful agency over how data derived from them is used, shared, and repurposed.

particularly when data sets are combined, commercialized, or deployed in AI systems. Consent must be informed, revocable where possible, and supported by accessible grievance mechanisms. The framework must also build accountability into the system, clear standards for access, independent oversight, anonymization, and remedies when misuse occurs. And critically, openness must be tied to domestic capacity building. Data sovereignty has little meaning without adequate capacity. Public data should not simply circulate globally. It should strengthen local research institutions, startups, digital infrastructure, and technological expertise. Domestic digital law should prevail over foreign commitments. At the same time, none of this implies that countries should isolate themselves digitally. Cross -border data flows are essential to research collaboration, to trade, to financial systems, and technological innovation.

Digital ecosystems simply do not function in silos. However, enabling data to move across borders should not mean that countries give up the ability to regulate how that data serves their own development priorities. Interoperability should facilitate cooperation, not erode policy space. This balance between openness and sovereignty is already reflected in recent multilateral commitments. The G20 New Delhi Leaders Declaration in 2023 placed digital public infrastructure at the centre of inclusive growth and emphasised data for development, linking data governance with trust, security and domestic capacity building. The message was clear. Data must support development, not undermine regulatory accountability. Similarly, the Global Digital Compact adopted by the United Nations calls for safe and transparent trustworthy data governance, stronger digital capacity in developing countries.

and international cooperation that respects national regulatory frameworks. Together, these signals suggest that the emerging consensus is not about unrestricted flows or digital isolation, but about structured openness where innovation and cooperation coexist with sovereignty and institutional strength. If we widen the lens, what emerges is not a contest between openness and sovereignty, but a conversation about how different regions are navigating that balance. The European Union has demonstrated how strong regulatory architecture, through instruments such as data protection and digital market rules, can shape global standards. India, by contrast, has shown how digital public infrastructure can scale inclusion at population level. India is putting innovation ahead of regulation. These are not competing models. They are complementary experiments in digitalization.

Global governance and increasingly the global south is not merely observing this evolution, it is participating in it. India’s experience with IndiaStack illustrates what this participation can look like. By building interoperable layers, digital identity through Aadhaar, real -time payments through UPI, document exchange through DigiLocker, India has created a public digital backbone that supports innovation while remaining accessible and adaptable. Crucially, this architecture has been offered as a template for other developing countries, seeking scalable and affordable digital solutions. In doing so, India has reframed digital infrastructure not as proprietary leverage but as a developmental public good. Of course, much remains to be done. Questions of data protection enforcement, AI governance, cyber security, resilience and equitable access require sustained attention.

but the direction is clear India is not approaching the digital future as a passive market, it is shaping it as an architect as conversations advance from G20 to the Global Digital Compact and now through initiatives such as this India AI Impact Summit the emphasis is increasingly on responsible innovation capacity building and inclusive growth. Our trade agreements must not promote digital dependency or virtual vassalage. We must emerge as a digital sovereign empowered to protect our own giants and capture the wealth generated by our own data. Friends, the task before us is not to choose between openness and control but to design systems that honour both. If we succeed open data will not be a source of vulnerability but of empowerment and in that journey India alongside partners such as the EU and our fellow countries of the global south has the opportunity not merely to catch up but to help define the rules of a fairer digital order rather than subject ourselves or submit to subaltern status under a new extractive digital at large.

Okay, Raj?

Thank you, Shashi, for setting the tone for this. So we, as you can see, of course, Prime Minister Hacker and Humphrey and others are sitting here, and then after hearing this speech, Humphrey is remarking, Prime Minister, if we start to believe what Shashi Tharoor is saying, we may end up in a situation where governments begin doing what they must rather than what they prefer. We may be entering a new administrative era. And Prime Minister replies, Hacker says, good. And so Humphrey replies, terrifying. And Bernard says, Prime Minister, shall we remain in this panel? They are about to discuss statutory mandates. And Humphrey replies, I do hope it’s only exploratory. With that word, may I now invite our distinguished panelist, Ms. Rama Vedashree.

May I request all our panelists to keep it for three to four minutes so that we can hopefully have another round. So, Ms. Vedashree, you’ve had a long and distinguished career pretty much designing these things and providing leadership. So my question to you is that when and how did the idea behind the national data sharing and accessibility policy and the open government data platform essentially germinate? Take us through the journey and also the challenges that you face through this.

Sorry, I’m here.

I’m sorry I didn’t notice that. Take us through the journey and help us understand how the concept moved from its formative face to a reality. Thank you.

So actually this open data moment was… It was a global movement. In India, actually it was former colleagues of cabinet colleagues of Mr. Tharoor, Mr. Kapil Sibal and Mr. Sachin Pilot when they were in the ministry. It started then and then the national data sharing access policy, I think around 2012. And industry also contributed to that entire draft and that policy. At that point of time, I think the entire focus was on opening up government data. And maybe some development data. So it was mainly government data. And then the data .gov .in platform came. What we need to take stock of is that entire open data movement and our own NDCEP policy and data .gov .in platform was built where to open up government data probably for research and other policy making.

Innovation, I mean opening up this for innovation, and I’ll start to. was not really a primary objective because it was in the pre -startups era and the pre -AI era. And that’s where I think now we need to really revisit that and make sure that we’re just not locked down by the old paradigm of open data because right now you need open data but which is also AI -ready open data, which is extremely important because when you look at an LLM or any other small language models, there are end users, there are professionals, there are researchers, everybody using that, prompting it, and they’re expecting the data. So in the past when the open data movement started, I think we were happy if government opened up by giving us a PDF or CSV file and we would figure out, download, put it in a spreadsheet and do our analysis.

Whereas now, the data needs to be truly open. always available and most importantly I think metadata and its standards are extremely critical for the interoperability of the data and we need to revisit how different segment of users of this open data are going to consume that data. We are now also in the, nobody wants to download and do something offline, right? They want to be able to consume the data through APIs and through apps and then of course the entire AI systems and we need to make open data available where not only end users like you and me can consume but both apps and AI systems can consume. I think that is where we have a challenge and having spent 35 plus years in the industry, I beg to submit that just opening up government data is not enough.

There is a lot of institutional data which is getting locked and siloed. I would like to call it daft data because nobody is using them. even in commercial enterprises and with regulators and nodal organizations like CERT. So when you look at cybersecurity startups, they really don’t care about what is there on open data .gov .in. They need a lot of data which is there with the nodal institutions of government. Similarly, with regulators, fintechs want data that is residing with NPCI. So we need to look at how do we open up this data.

Thank you so much, Vedashree. That was very concise and even compelling. Especially coming from a regulatory standpoint. May I invite Ms. Irina Ghosh, the Managing Director of Anthropic India. So Ms. Ghosh, my question to you is that as Anthropic deepens its collaboration and presence across India, could open data sharing frameworks help drive trust -first innovation and development in the Indian AI space? Is this relevant at all for AI developers such as Anthropic in making the AI models more secure, trustworthy, and well -suited for complex dynamic and regulatory development? And rapidly evolving markets such as India, Thank you so much.

It’s indeed a pleasure and a complete honor. And I really love the analogy and the follow up thereafter as well. Let me first begin by saying that I think all of us totally agree that AI for India is a generational opportunity in the context of the data, the demographics and the culture. Having said that, it’s not the question that is it the AI moment for India? It’s a question do we trust and do we want to make it the AI moment for India? And trust for all of us needs to be a verifiable outcome. Do we trust the data that we are putting in every click, every transaction, every decision which is triggered by the AI?

Is there an invisible filter or are we trusting it? So two parts to that in my mind. One is the data that we are collecting. How are we using it and making it available? For many more experimentations and innovations there. If the model is only being built on a western data or for financial institutions which is serving a different segment or a sector. it won’t be communal useful for all of us so few things that we need to do is make it contextual to the local language and the domain in the local language legal agriculture for the languages which are there in India that’s the first thing now how do we ensure that it goes across at scale that’s the second there are three things that we are doing first of all we are doing an economic impact survey index survey by which we are ensuring that we are really making data available for the way people are using it in India and a big round of applause to everybody out here because the highest usage of CLOD which is the tool anthropic users is from India so we have a great way of knowing as to what people are doing and we share it completely contextually as to what people are using it for that’s the first the second people will want to use the data and the context collectively don’t do it once but don’t rewrite code the analogy to that I would say is that when you had a mobile phone world you did not want to have a charger for different mobiles right the universal connector came across that solved all the problems so when you create, look at a farmer when he is wanting to use things there are 3 -4 kinds of data, the market index, the soil data, the irrigation data, if you try to pull in data every time and make it work, it’s gonna fail so a model context protocol, MCP as we call it was created by Anthropic in 2024 and we put it across to the Linux community, anybody and everybody can use it so that once you create an AI layer on top of that, people can pull that data why is it contextual for India?

There is a lot of data which is lying across in agriculture, health, education and like Rama called out in institutions and we are working along with the collaborations of all the players, Google, Anthropic, Microsoft, every single one everybody else put together and when the Honourable Prime Minister called out the manifesto, we are ensuring that we make data transparently available. We are also committing that we will build it across for use cases in the sectors which mean the most to India so that we emerge and make it the AI moment for India.

Thank you, Ms. Gose, for really giving that perspective. Now may I invite Mr. Cyril Shroff who is of course the convener of this panel but also managing, founding, managing partner of Cyril Amachal Mangaldas and the benefactor of the Cyril Shroff Centre for AI Law and Regulation. Mr. Shroff, in your view, might a clearer regulatory framework be necessary to ensure more consistent, effective and systemic data sharing by government bodies? The clarity that we need from you is that the role that a regulatory framework can play in institutionalizing incentives and accountability and putting in place initiatives. The role that a regulatory framework can play in institutionalizing incentives and accountability and putting in place initiatives. The role that a regulatory framework can play in institutionalizing incentives and accountability The role that a regulatory framework can play in institutionalizing incentives and accountability The role that a regulatory framework can play in institutionalizing incentives and accountability and putting in place initiatives.

courts. I say this because of the fact that most of the time, lawyers come to the party very late. The technology is so fast and things get done and when shit hits the roof, to put it bluntly, lawyers are asked to clean it up. Should we do it differently?

innovation and regulation, and the regulation here is intended actually to create the foundation stone for innovation. If data was systematically available in a usable format, AI -ready format, that would actually spark a lot of innovation and create the foundation for it. So I think that the short answer, as I said, is yes. And I think just to build on Dr. Tharoor’s point on data sovereignty, I think the Prime Minister said it well. And I think he had probably the saying, when he actually said that, and I think as India we need to assert that right. All the data is largely in the global south, and all the companies and the private sector and the usage is largely in the global north.

I think we need to assert ourselves. I think that’s exactly what Dr. Tharoor said, and I’m a great fan of that. and it partly kind of explains why in a personal philanthropy level I created this center because lawyers come late to the party but some lawyers don’t. So I think this is what I expect from your center Raj. So I’ll stop there. I think we have a lot to come.

Lawyers have another quality. Put the blame on somebody else when things are not happening as much. That is known as good management. Thank you so much Cyril for that because I think it’s important for us to recognize that while we are indeed attempting to frame regulation we also should not stifle growth and innovation because that’s the biggest death knell that we can sound towards a lot of entrepreneurship that’s emerging. May I now invite Dr. Sasmit Patra. Sasmit you are a distinguished member of parliament and of course you have straddled across the world of policy making and even academia. How can greater availability of reliable public data lead to stronger evidence -based policymaking and more efficiently delivering public goods?

In fact, the real question is the criticality of data in identifying relevant areas of policy intervention by the state for designing public policy instruments and frameworks so that targeting to the relevant stakeholders. How do we do that?

Thank you, Raj. It’s a very important question because I’ll take it in two parts. The first part is whether data is important to policymaking. Yes, it’s a no -brainer. Second is, in a federal structure, the problem is data is in silos. Let’s say I come from the state of Odisha. So the data of our farmers in our local Kalia Yojana would be in a different format and probably kept differently than probably the PM Kisan data that is there by the federal government. secondly how can this be useful I’ll come to the second part let’s say crop loss Pradhan Mantri Fasal Bhima Yojana is something which is a crop loss provision that is given for reimbursement or compensation for crop losses for farmers in this scenario what happens is if the data is readily available with the government the government can predict that over the next 1 to 2 years which are the districts which are the taluks which are the blocks which are the panchayats where crop losses have been happening over a period of time so predictively AI can bring about solutions to probably A. try to find out the reasons for crop loss B. try to mitigate the losses C. try to strengthen the farmers for crop diversification and D. try to generate a new form of mitigation plan that can be implemented by the government in those in order to do that you need data without that data it is not possible I’ll come to the last part where the government will actually have a problem it’s a political question The political question And I’ll play the politician here The reason is When the government says I’m going to share data It’s a data of 140 billion 1 .4 billion people right How many of you sitting in this room here Are willing to share your data Through the government Anonymize the data That’s the question I’m trying to put to you Let’s say tomorrow government Comes up with a regulation and says I want to share that data A trust and verifiable data My citizens data Not the farmers data Who is nameless and faceless At Bharat Mandapam The movers and shakers of Delhi data Is going to be now released For training of LLMs and micro LLMs Are you happy sharing that data That’s where the catch is That’s the political question The regulatory question is Yes there has to be the data The policy question is The data is needed for better policies But see as citizenry The data is needed for better policies How many of you sitting in this room Are comfortable sharing the UPI transaction that you do.

That, even if in an anonymized question, will always remain. So the answer starts with you and ends with you as a citizen.

Thank you so much, Sir Sussman, for that very important question. I think it’s important to recognize that the heart of it is about to what extent citizens are prepared to trust the government. And trust factor becomes critical here. Let me quickly move to Miss Vedashree. We’re doing very well on time, so thank you for all the panelists to respond with short responses. So Miss Vedashree, why in your view have proposals such as India data accessibility and use policy and the national data governance framework policy well -intended policies, but have not really moved forward? Why it has been the case that these government interventions being only at the policy level, lack regulatory enforcement?

So the first thing, I think, we need supply -demand gap assessment, because maybe government is opening up or throwing up some data on the data or the government in platform. who are the users who are consuming it and the ministries who are anyway overloaded with so much work and if they need to manage this and regularly submit all the data sets in an open format, they need to see what will come out of it, which means we need to tie it up with researchers, we need to tie it up with inventors and innovators. I think that did not happen so far. Whereas now if you really look at an AI -ready data, sets, she talked about there are open standards so that interoperability, she talked about the protocol, MCP protocol.

I think we now need to look at what data sets are needed for research, which could be academia and research students and for industry, which could be all the startups. Unless we map that and revisit what is the necessary policy and government data will be useful, development data, which even World Bank. throws up a lot of data in the open data sets. Those development data is also equally important for policy making. But if you’re looking at it, opening up data repositories, dark data as I call it, for innovation purposes, I think we need to look at how do we open up commercial data in a secure, anonymized way. There have been some steps, sir. For example, the payment systems directive in UK.

Now EU, who’s always been about extreme of protecting data, is now talking of FIDA, which is a financial data access, where they’re saying at a sectoral level, how do we open up the data access? Payment systems directive and open banking initiative was that. Similarly, healthcare data, hopefully the Aishman Bharat mission will open up. So I think we need to look at the supply -demand gap, what data will be consumed by which segment of users, and open up those data sets. Otherwise, I don’t think we will move anywhere.

Thank you so much, Ms. Vedashree. I haven’t forgotten you, Arun. You’re, of course, our own. Arun, of course, is a partner at Cerebral Medicine, Mangadas. If India were to move towards a more structured legal framework for open data sharing, which core principles and safeguards could shape its design? Is it all about design thinking so that government bodies requiring them to share the data have aggregated data sets on a free, paid or restricted basis with voluntary private participation on the supply side and, of course, safeguards to prevent misuse?

Thanks, Raj. What I lack in the eminence or erudition of my fellow panelists, I will try to compensate very inadequately with a certain radicalism. Not the radicalism of rhetoric, but the radicalism of making bold suggestions as to the minimum viable proposition of a sustainable open data. Thank you. And by positing that the lack of open data sharing that several of the key panellists, including the keynote, have called out has arisen due to a lack of a durable legal architecture. Today, in India, despite having a Digital Personal Data Protection Act, episodic intermediary regulation, as well as several policy and practical initiatives on the sharing of non -personal data, we do not, as the world’s largest democracy, have a clear identified anonymisation standard, clear identified public data interchange standards.

We do not have a clear recognised purpose for the processing of open public data sets for public good and public improvement. This means that any initiative, particularly large complex initiatives like large language models and their deployment, which are multi -decadal, multi -billionaire. These multi -billion dollar investments are open to the travails of both judicial storms. executive weather patterns and perhaps most importantly legislative climate change a government official who creates an open data repository has to risk that in 5 years his action may not only be frowned upon but be downright illegal a founder betting his life on creating the next generation of open data architecture and applications has to risk at some point that his business becomes fundamentally unviable I submit to you that absent these 4 key important elements which work coherently not only with existing architecture but also the constitutional principles which have been laid out in the Puttaswamy judgment which continue to pervade our democracy until that architecture is enacted in legal legislative form in a way that does not rub up against the various pieces of isolated sectoral regulation we have across individual regulators we will not have a sustainable open data ecosystem Thank you.

Thank you so much, Arun. That was fantastic. Spoke like a true lawyer. Cyril, quickly, we have a few minutes left, and we have concluding remarks as well. From your vantage point, how can greater availability of reliable public data influence investor confidence, efficiency of markets, and long -term economic growth? In many ways, this is also a moment for India to showcase its potential for attracting investors to believe in both the government and their investments to have the right results.

I’m going to answer your question with an analogy. One of the hats that I wear is also as a capital markets lawyer. And I’ve seen how the growth of India’s capital markets from a very restricted, basic kind of, you know, at a very fundamental level to what it is today as one of the most vibrant capital markets in the world, last year. India had 25 % of all global IPOs even more than the U.S. And why did that happen? That happened for a variety of commercial reasons, but also the fact that we have a very vibrant capital market regulatory system in place. It has taken 25, 30 years to get us to this point. But a lot of it is about having regulatory clarity.

It is about having the right enforcement. It is about having good accounting standards. It is about just uniformity in regulatory language that is used, at least the same vocabulary, which is if you can just substitute the word capital market by the data and the digital world, I think you get to the same answer. So I think the answer lies, therefore, in that if you want to create trust in the community, if you want the multibillion -dollar investments, if you want data centers to be set up here, if you want us as a country to move from a – from a services -based tech sector to a product -based tech sector and – it may be different parts of the digital world, I think you first have to create trust, and a trust cannot happen without transparent information and a reliable legal and policy system.

Now, one of the things that we get periodically hit on the head with is the fact that your dispute resolution system is too slow, it takes 30 years to enforce a contract, blah, blah, blah, and something which we take disproportionate stick for. But I think a lot of it ultimately comes down to can you trust your legal system. And I think the answer is if we are able to create that right regulatory policy and enforcement framework for this, which kind of answers your question, I think we would have solved it. It’s not going to happen otherwise. There’s no point having a law which you can’t enforce.

All right. Thank you so much, Cyril. I have enough sign language indications to say that we have another 10, 12 minutes, so I am proceeding forward. Shashmit, quickly to you, are there any geopolitical concerns that need to be addressed if open data sharing practices by the government are to be scaled up in India? Should that be something that we need to be concerned, especially because you’re sitting in the parliament and there are, of course, opposition parties. really coming forward to question and challenge the government on this matter as well.

You know, in fact, when recently the US -India trade deal happened, then you had a lot of energy being seen in the parliament and outside. So sharing of data and the method by which we share data is of course a geopolitical concern for the country. So maybe we can look at data, as Madam just said, that one is the data that is for the public good and the humanity. The second data is restricted and probably national security. And the third data is something that can be monetized and commercially useful. So therefore, instead of probably putting the entire data set within one silo, we can probably look at the usage and thereafter tag them to the respective so that the multi -billion dollar innovator also benefits, the regulators also benefit, the citizenry also benefits, and finally the policies for the farmers, the Anganwadi workers, the ASHA workers, also get done.

Last point, and I just want to put that on record because I’m on the… the Parliamentary Oversight Committee on Communications and IT, and Dr. Tharoor was the earlier chairman and my distinguished colleague there as earlier. One of the critical areas that we at least are discussing and debating is not a very hard, strong EU AI Act. I don’t think that’s happening anytime soon. We’ll have a regulatory framework, and the key word is soft -touch regulation. Where does that take us has to be seen.

Thank you so much, Sushmit. We’re very fortunate to have both Shashi Tharoor and Sushmit Bhattabhai, former chair of the same Parliamentary Committee, and Sushmit now a member. May I quickly invite Ms. Vedashree to do a one -liner concluding response, especially as you look at from this vantage standpoint, how do you think the future is going to evolve, especially in the light of India wanting to play a global thought leadership role? I think this summit is demonstrating that. Our aspirations of positioning ourselves.

as, you know, the AI leader of the world. We are working towards that. So I would say when you link it to the topic, I think we need a very concerted data strategy at a government level. There were some efforts when the personal data protection bill was being debated upon. There was also a parallel one around non -personal data framework. So I think we need a national level data strategy because now we need to look at it from the current to the next five years. How do you open up? Sir talked about, you know, that data needs to be different segments. I also believe that we cannot have one centralized open data repository. Data needs to be federated.

We also need to think through, along with the sectoral regulators, what will be the sectoral data opening up policies because that’s where a lot of… data that can be monetized and innovation can happen. So we need to look at that at a sectoral level and at a government level and how do we create this federated open data strategy which is ARAD.

Thank you so much. Your one -liner, Cyril, one -liner about what should India be doing as we look at the future?

Not copying the West.

Good one, good one. Alright. May I invite Ms. Irina Gross for you to especially from the standpoint of anthropic but also the private sector which is expecting a huge presence in India.

Yeah, I think the last mile between making AI real is the diffusion which has to happen between the frontier firm which is creating the model and the person on the last mile who is needing it and that’s the thread of trust. Now the thread of trust needs to be woven by the contextual data in the context of India and ensuring that we are making it both open, accessible and ensuring that everybody is contributing to that grid.

Thank you so much, Arun. Over to you. Your one -liner.

The absence of a legal framework goes from being an inconvenience to an impediment in the development of a sustainable data economy We are at the point where India’s existing regulatory framework is making that transition Thank you so much

We have now come to almost the end of this panel but we have a very distinguished panelist Ms. Asha Jadeja Motwani She has been silently and quietly listening to everybody but she is at the heart of India -US relations but also somebody who has been a venture capitalist investing in tech companies, innovation as well as AI in India and in the United States She has been working hard to build that relationship but also has been a benefactor just as Cyril established the Center for AI Law and Regulation Ms. Motwani established an endowment at our university where we have established India’s first Motwani Jadeja and Ms. Asha Jadeja Institute for American Studies So may I invite you to share some reflections having been part of the Global AI Summit and of course this particular panel Over to you Ms.

Jadeja

Yeah, thank you, Raj, for inviting me. And so the one thing that I want to, you know, stress heavily is that, look, we are built on an American stack, you know, and we want to make sure that, you know, one of the things I heard in the AI summit was this question of sort of what if, you know, what if America at some point becomes more of a hostile entity and pulls its APIs? Will we be stuck? You know, will we be in a situation where we don’t know how to handle it? I think that question is something that we must think about and know how to deal with it if it happens. But I don’t think it’s likely to happen.

We will actually have to make a decision and say that we have consciously chosen to be on the American stack. From the, you know, from the chip level all the way to the top. and so if we consciously make that decision then at a policy level and even probably even at the legal level what you guys will need to figure out is that if we have decided to work with the Americans on this and put our eggs in that basket then have a joint regulatory framework so that we are never conflicting with them number one number two also to make sure that we don’t get a situation where we are holding back data because remember the AI revolution is all about training the new models training these new entities that are going to be a doctor in our pocket for training those things we need to make sure that our data which is the health data for example Indian health data is open and accessible to those in the west who are developing these programs these models so it’s critical to know that it’s a fine balance this is not like the internet business this is not like the internet And, you know, we had to worry about, you know, who is going to do what with that?

Are they going to pump ads at us? This time, it’s much more about what are these things going to give back to us once they have that data. So it’s a tricky balance. And I think we will need to make that decision about do we trust the Americans and do we trust the American stack? And if so, how do we proactively work with them so that their hands are also tied just like our hands would be tied?

Thank you so much, Mashaji. In fact, it’s also important for us to recognize that it is also the bedrock of TUS is based upon should India be working with democracies? Should India be working with countries with more shared values and societies which largely, you know, recognize the importance of rule of law and democratic institutions? And that’s where the India -U.S. relations lie. we have of course had a wonderful panel discussion but as a professor I will not let our panelists leave without an audience question we have a few minutes left so I request if anybody, let’s have the mic to the lady here in the second row keep it very short, I’m going to collect three questions and have our speakers respond

thank you for giving the opportunity, I’ve been dying to ask this there have been a lot of sessions where we have been talking about having a regulatory framework on AI and having independent assurances but the way things are right now even the auditors, the regulators etc are heavily reliant on AI so who would be watching the watchers?

Good one, good one, Cyril you who would be watching the watchers alright, next question the standing people, let’s give the standing person mic here they’ve been standing for long

BK Patnaik (Audience Member 2):

I’m BK Patnaik from Orissa I am asking the question to Mr. Patra that you can ask in Odisha itself anyway Mr. Patra what he has told that he will change the farmers in India with the AI data will it be successful I am asking to Mr. Patra

last question there and another to Mr. Tharoor one one one just you can raise your voice go ahead

so how you got a measure that started for men exclusively it’s called Dora Health exclusively for men now my concern that I want to share on this panel where we are talking about regulatory framework for AI is that I have seen that it’s very difficult to get information for men specifically to work on techniques that can help them governments have failed to search on it companies have failed to search on it even third party cookies that are shared don’t really look into this specific aspect and men’s mental health is just one of them so what do you recommend that the government would is is still as a regulatory framework is still as a regulatory framework is still as a regulatory framework is still as a regulatory framework is still as a regulatory framework is still as a regulatory framework and and and is still as a regulatory framework ensure that such precise data is given into the right hands

Got you. Let’s have Mr. Vedashree answer that question. But Cyril and Sasmitia.

Yes, yes, yes. So I think you raise the right question. So this is where this, we are not actually discussing regulation of AI. This was around open data. Just to clarify. So I think the sectoral data that when I talked about, when it’s healthcare data, you’re talking about mental health data. But it’s very rarely that personally identifiable data will ever be opened up. I don’t think it will ever be opened up. And that is important. But this is where I think, for example, in Germany, there is some healthcare act in which they’ve given a provision where patients can choose to share, ask their healthcare institution to share some specific data. That maybe I will, let’s say I’ve had some critical illness.

I’m willing to share everything anonymized so that it goes for research. So I think we need some progressive regulations and also education and awareness. Only regulation will not open up this data.

If I may just add something to it. The reason I said regulation is because I’m actually working on an AI pilot system right now where it can analyze the charts that take place between a man and a confidant that works for us. So after analyzing this chart, we have a very procured list which is non -sensitive data that can be shared. So what I’m asking is as a government, what is the regulatory framework that can be put into place so that this non -sensitive data can be shared for research?

So there is some discussion going on. The guidelines came from Ministry of Electronics and IT around AI. So we can expect some movement around it. This could be taken offline as well.

Dr. Patnaik, lovely shades. As far as … AI is concerned, I think yesterday when the Honorable Prime Minister inaugurated, he said, we are looking at humanist AI. I think that is the cornerstone, inclusive and humanist. If it doesn’t solve the problems of the farmers, doesn’t solve the problem of the healthcare workers, doesn’t solve the problems of the tribals in the state of Odisha and elsewhere, then how is AI going to benefit humanity? So therefore, the Indian AI thought leadership, so to say, comes from the Indian concept of Vasudeva Kutumbakam. The whole world is one family. We are humanist, we are inclusive, we want our AI to help everyone.

Thank you so much, Sasmit. Over to you, Sir.

So I’ll take your question. The question was who is going to mind the watchers? Or who is going to watch the watchers? And I think the answer lies in our constitution. The answer has been there for the last 75 years, the courts and the rule of law. We are, actually, I think we have the best rule of law system in the world. Earlier I used to give a credence to the United States for that but everything that has happened in the last 18 months has shown that that’s not true and their legal system can be arm twisted, can be pressured lawyers can be pressured, all of that that can never happen in India we actually have a much better democracy and even though it may be clumsy in the way we sometimes go about it and it may be frustrating, there is only one answer in India which is the courts and the second answer I think is ethics so one of the, and AI is going to need a completely different ethics code about biases, about so many things, it may not be law but it is something which the industry will have to evolve for itself there are number of topics around which and I think we are working on that in the center, so these two answers, these two themes will provide the answer to how are you going to regulate all of this, the second one is a bit more ambiguous because ethics conversations always are more amorphous but the courts like it or not finally they will be the one Putuswami is an example of that and there are so many similar examples finally the course always comes through there hai andher nahi hai

thank you so much sir we started with the first word by Shashi Tarur we will have the last word by Shashi Tarur

that’s been a fascinating discussion I think some of the questions raised and perhaps the ones that weren’t raised are already pointing the way to some of the further areas we need to converse about but we also have to be very realistic and anchored in what we are talking about when our friend from Orissa asked that question about agriculture I like Sasmit Patra’s answer as an aspirational answer but I am very conscious that even as the finance minister announces a special budget provision for AI in agriculture that the vast majority of our farmers can’t afford tractors can’t afford tillers don’t have pumps don’t have guaranteed sources of water and in many cases no 24 hours of electricity and in those circumstances AI how is it going to be applied, how many farmers will it reach and how in how many ways will it transform agriculture.

I think humanist agriculture, humanist AI is an laudable goal but we have to relate it to the reality of our own people and the circumstances in which we’re in. I mean I would love to agree with Siddharth Shroff on pretty much everything he says but with our legal system we face for example the undoubted fact that your judiciary needs AI to begin with. I mean you’ve got 5 million pending cases in this country. How can we celebrate the rule of law?

50 million.

50 million. So how can we celebrate the rule of law when justice delayed and that old cliche is justice denied. So again let’s anchor all this into the real world. When you speak of men’s health or anybody’s mental health for that matter, it seems to me you’re touching on an extremely important issue. But a lot of this stuff is the circumstances that the person is confiding into a doctor, in your case, into a chat, how much of that can be anonymized truly effectively, how much of it can be traced back, how much of it can cause confidentiality breaches. The purpose of health data aggregation ought to be to solve similar problems for other people. In other words, let’s say doctors in the West may have access to 10 instances of a rare disease, and in India there may be 1 ,000 instances of the same disease.

So if we had AI data in India that aggregated all that, then certainly the West might have the scientific technology to research it and come up with a cure or a better cure or whatever that can be applied in India. But for all that to happen, we need regulations. We need to figure out if there’s monetization, who benefits, if the data is given on what. What terms, how do we have a law that says what will come back to us, etc., etc., etc. It would be absurd if those 1 ,000 Indian cases added to the 10 Western cases to create a proprietary. fighting into a doctor, in your case, into a chat. How much of that can be anonymized truly effectively?

How much of it can be traced back? How much of it can cause confidentiality breaches? The purpose of health data aggregation ought to be to solve similar problems for other people. In other words, let’s say doctors in the West may have access to 10 instances of a rare disease, and in India there may be 1 ,000 instances of the same disease. So if we had AI data in India that aggregated all that, then certainly the West might have the scientific technology to research it and come up with a cure or a better cure or whatever that can be applied in India. But for all that to happen, we need regulations. We need to figure out if there’s monetization, who benefits, if the data is given on what terms, how do we have a law that says what will come back to us, et cetera, et cetera, et cetera.

It will be absurd if those 1 ,000 Indian cases, I added to the 10 Western cases, to create a proprietary… AI software that the 10 ,000 people here can’t afford to benefit from. So we need to create models. That’s the point that I was trying to make when I talked about the digital Raj, but we have here a moderating Raj on action, so I better stop. Thank you all for listening. Thank you so much. Thank you.

Thank you, Arun Prabhu, Irina Gross, Asha Jaliza Motwani, Shashi Tarur, Cyril Shroff, Saswit Batra and Ms. Vedashree. I want to particularly thank Shashi who came for the keynote address and was planning to leave in 10 minutes, but he stayed through the entire panel. So give the entire panelist and Shashi a big round of applause. Thank you to Asha who agreed this morning. So thank you very much.