Panel Discussion AI in Healthcare India AI Impact Summit

20 Feb 2026 17:00h - 18:00h

Session at a glance

Summary

This discussion focused on the opportunities and challenges of implementing AI in healthcare systems, particularly in India and other emerging markets. The panel featured Dr. Sabine Kapasi, Chris Ciauri (Managing Director at Anthropic), and Dr. Aditya Yad (India Relations Advisor at Invalude), who explored how AI can transform healthcare delivery in low- and medium-income countries.

Chris Ciauri highlighted two major areas where AI can create significant impact: reducing administrative burden in developed countries like the US, where doctors spend only 30% of their time on patient care, and improving healthcare access in India, where primary care visits average just two minutes. Anthropic has opened operations in Bengaluru and trained their Claude model on 12 Indic languages to address multilingual barriers in healthcare delivery. The company emphasizes safety in healthcare AI applications, ensuring their models acknowledge uncertainty rather than providing overconfident but potentially incorrect responses.

Dr. Aditya Yad discussed Switzerland’s perspective, noting the country’s leadership in global innovation and its expensive but high-quality healthcare system. He explained how AI is being integrated into drug discovery, manufacturing, and clinical processes to reduce costs and improve efficiency. The recent Switzerland-India free trade agreement includes a commitment to invest $100 billion in India over 15 years, with healthcare as a key sector.

The panelists identified several promising use cases for AI in healthcare, including administrative task automation, drug discovery acceleration, diagnostic improvements, and workflow optimization. They emphasized that AI should serve as a preparation tool for clinicians rather than replacing medical judgment. Key challenges discussed included building trust around medical data usage, training healthcare workforces for AI adoption, and ensuring equitable access to AI-powered healthcare solutions across different economic contexts.

Keypoints

Major Discussion Points:

– AI’s transformative potential in healthcare with emphasis on safety: The panelists discussed how AI can revolutionize healthcare delivery, from reducing administrative burden (70% of clinicians’ time in the US) to improving access in countries like India, while emphasizing the critical need for safety-first approaches that acknowledge uncertainty rather than providing overconfident but potentially wrong answers.

– Geographic opportunities and challenges in AI healthcare adoption: The conversation explored how different regions face distinct healthcare challenges – administrative burden in developed countries versus access issues in India and the Global South – and how AI solutions must be tailored accordingly, including multilingual capabilities and culturally appropriate implementations.

– Drug discovery and manufacturing transformation through AI: Participants highlighted how AI is revolutionizing pharmaceutical development, reducing drug development timelines from eight weeks to eight hours in some cases, and enabling more efficient biomanufacturing processes with better yields and lower costs.

– Workforce enablement and the human-AI collaboration model: The discussion emphasized that AI should serve as a preparation tool while clinicians retain judgment responsibilities, focusing on how to train healthcare workers to effectively leverage AI while maintaining appropriate skepticism and professional oversight.

– Data trust, policy considerations, and market dynamics: The conversation addressed critical concerns around medical data privacy, the need for public trust in AI systems, and the evolving landscape between large language models versus smaller, targeted healthcare applications.

Overall Purpose:

The discussion aimed to explore near-term opportunities for AI adoption in healthcare, particularly in India and other emerging markets, while identifying strategies to strengthen healthcare systems for real-world AI implementation over the next 3-5 years.

Overall Tone:

The tone was consistently optimistic and collaborative throughout, with participants expressing genuine excitement about AI’s potential to transform healthcare globally. The conversation maintained a professional yet accessible atmosphere, with speakers acknowledging both opportunities and challenges while emphasizing the importance of responsible AI development. There was notable mutual respect between the technologist, clinician, and policy perspectives, creating a balanced and constructive dialogue that remained forward-looking and solution-oriented from start to finish.

Speakers

Speakers from the provided list:

– Dr. Sabine Kapasi: Clinician/surgeon, discussion moderator. Has experience practicing as a surgeon and seeing 200 patients per day in OPD (outpatient department).

– Chris Ciauri: Managing Director at Anthropic, leads global expansion across EMEA, APAC and Latin America. Has over 25 years of experience scaling SaaS and cloud businesses, including senior leadership roles at Salesforce, Google Cloud, and was previously CEO of Unilever. Brings expertise in enterprise AI adoption and national technology growth.

– Dr. Aditya Yad: India Relations Advisor at Invalude (innovation and investment promotion agency of Canton Broad, Switzerland). Based in Lausanne, he is a biotechnologist who works at the intersection of tech, investment, and innovation. He facilitates Switzerland-India collaboration, supports startups, and enables market entry for Indian companies into Swiss and European innovation ecosystems. Also serves as a policymaker and legislator in Switzerland.

Additional speakers:

– No additional speakers were identified beyond those in the provided speakers names list.

Full session report

This comprehensive discussion on AI in healthcare brought together diverse perspectives from technology, clinical practice, and policy-making to explore the transformative potential of artificial intelligence in healthcare systems, particularly focusing on opportunities in India and other emerging markets. The panel featured Dr. Sabine Kapasi, a surgeon with extensive experience in resource-constrained healthcare settings; Chris Ciauri, Managing Director at Anthropic with previous leadership roles at Salesforce and Google Cloud; and Dr. Aditya Yad, a biotechnologist, legislator, and policy advisor facilitating Switzerland-India healthcare collaborations. This session represented the final day of a multi-day event, with the discussion focusing on near-term opportunities over the next 3-5 years.

The Global Healthcare AI Landscape: Divergent Challenges, Convergent Solutions

The conversation began with a striking revelation about the fundamentally different healthcare challenges facing developed and developing nations. Chris Ciauri highlighted that in the United States, clinicians spend only 30% of their time on actual patient care, with the remaining 70% consumed by administrative tasks and paperwork—representing a massive efficiency problem. This administrative burden contrasts sharply with the access challenges in countries like India, where healthcare providers face overwhelming patient volumes. Dr. Kapasi shared her experience of seeing 200 patients per day in her OPD, illustrating the scale of demand that healthcare systems must manage.

This dichotomy established a crucial framework for understanding how AI solutions must be tailored to address region-specific challenges rather than adopting a one-size-fits-all approach. In developed markets, AI’s primary value proposition lies in reducing administrative burden and freeing up clinician time for patient care. In emerging markets, the focus shifts to improving access, extending healthcare reach, and enabling more efficient care delivery within existing resource constraints.

India’s Emergence as a Global AI Healthcare Leader

Perhaps the most surprising revelation of the discussion was India’s position as a global leader in AI adoption. Chris Ciauri disclosed that India has the highest adoption rate of Anthropic’s Claude AI model outside the United States, ranking second globally in usage. More remarkably, usage and revenue from India had doubled in just four months, demonstrating not only high adoption but accelerating growth. This challenges conventional narratives about technology adoption patterns and positions India not as a recipient of Western innovation but as a leader driving global AI implementation.

This leadership position is underpinned by India’s remarkable digital infrastructure achievements. The country has built what Ciauri described as “a digital healthcare system that’s the envy of the world,” providing an excellent foundation for AI implementation. Combined with smartphone ownership rates approaching 90% in urban areas and 75% in rural regions, India possesses the digital infrastructure necessary for widespread AI healthcare deployment.

Recognising this potential, Anthropic has established operations in Bengaluru and specifically trained their Claude model on 12 Indic languages to address multilingual barriers in healthcare delivery. This investment reflects a strategic recognition that successful AI healthcare implementation in India could serve as a model for the entire Global South, potentially influencing how AI-driven healthcare evolves worldwide.

Safety-First Approach: Balancing Innovation with Responsibility

Throughout the discussion, safety emerged as a paramount concern, with Chris Ciauri emphasising that “AI can do a lot of good. It also can create a lot of harm if done carelessly.” This acknowledgement set a responsible tone that permeated the entire conversation, moving beyond typical technology optimism to address real risks and implementation challenges.

Anthropic’s approach to healthcare AI safety centres on several key principles. First, their models are designed to acknowledge uncertainty, freely using language like “I don’t know” and “I’m not certain” rather than providing overconfident but potentially incorrect responses. This approach proved decisive in their partnership with Banner Health in the United States, where the healthcare system specifically chose Claude because they wanted a model that would acknowledge uncertainty rather than display false confidence.

The safety framework also establishes clear boundaries between AI capabilities and human responsibility. As Ciauri articulated, “AI is for preparation. Clinicians are for judgment.” This delineation ensures that AI serves as a support tool for healthcare professionals rather than attempting to replace clinical decision-making. Additionally, Anthropic maintains strict data governance policies, with Chris emphasising that Claude will “never use someone’s patient data to train our models”—a crucial ethical boundary for healthcare applications.

Transforming Drug Discovery and Manufacturing

The discussion revealed how AI is revolutionising pharmaceutical development and manufacturing processes. Chris Ciauri shared compelling examples from partnerships with major pharmaceutical companies like Novo Nordisk and Sanofi, where AI has dramatically reduced drug development lifecycle times from eight weeks to eight hours for regulatory and administrative tasks.

Dr. Aditya Yad expanded on this theme from a manufacturing perspective, describing how AI is enabling more efficient biomanufacturing processes. Smaller, AI-controlled bioreactors can now produce high-quality pharmaceutical products with better yields than traditional large-scale manufacturing, challenging conventional assumptions about economies of scale. These AI-optimised systems continuously monitor and adjust production parameters, leading to improved quality control and reduced production costs.

This transformation is particularly relevant given Switzerland’s position as a global pharmaceutical hub. With approximately 1,700 healthcare and life sciences companies and research institutions operating in a country of just 9 million people, Switzerland has maintained its ranking as the world’s most innovative country for the past 15 years. Dr. Aditya also highlighted India’s national priority focus on biofoundry policy, demonstrating how both countries are positioning themselves at the forefront of AI-driven pharmaceutical innovation.

Economic Implications and Investment Opportunities

The discussion highlighted significant economic opportunities emerging from AI healthcare implementation. The recent Switzerland-India free trade agreement includes a commitment to invest $100 billion in India over the next 15 years, with healthcare identified as a key sector. This investment is expected to create one million direct jobs in India, demonstrating the scale of economic opportunity that AI healthcare represents.

Dr. Aditya Yad noted that in Switzerland, $2.5 billion was invested in startups in the previous year, with many companies successfully raising funds by demonstrating AI integration in their development strategies. This trend reflects how AI has become not just a technological tool but a fundamental component of business strategy and investor appeal in the healthcare sector.

However, the panellists also identified significant economic challenges, particularly around preventive care and screening programmes. Dr. Sabine Kapasi highlighted the fundamental problem that screening precedes felt need for healthcare—people are reluctant to pay for healthcare services when they don’t perceive an immediate problem. This creates a systemic bias towards treatment rather than prevention, despite the well-established principle that prevention is more cost-effective than cure.

Workforce Development and Human-AI Collaboration

The discussion extensively explored how healthcare workforces can be prepared for AI adoption while maintaining appropriate clinical oversight. Dr. Sabine Kapasi raised critical questions about training healthcare professionals to leverage AI tools while preventing over-reliance on AI recommendations for direct patient care.

The panellists agreed that successful AI implementation requires a balanced approach to workforce development. Healthcare professionals need education on AI capabilities and limitations, but this must be coupled with maintaining clinical scepticism and professional judgement. The goal is to enhance rather than replace human expertise, enabling healthcare workers to be more efficient and effective in their roles.

Dr. Aditya Yad described Switzerland’s government-funded programme targeting CEOs and leadership teams of healthcare companies through structured cohorts. This programme focuses on strategic AI implementation from the company’s inception rather than retrofitting AI into existing processes. However, he acknowledged the scale challenge of reaching 40,000 small and medium enterprises in Switzerland alone, highlighting the broader challenge of ensuring widespread AI adoption across diverse healthcare organisations.

Real-World Applications and Use Cases

The conversation was grounded in practical examples that illustrated AI’s potential impact. Dr. Sabine Kapasi shared a compelling clinical anecdote about treating a dengue patient in a remote region where diagnostic tests were more expensive than treatment drugs. This story highlighted how economic constraints often force healthcare providers to make treatment decisions based on clinical judgement rather than optimal diagnostic protocols—a situation where AI could potentially improve both diagnostic accuracy and cost-effectiveness.

Chris Ciauri provided concrete examples of AI applications already showing results. Banner Health’s use of Claude to summarise complex oncology reports demonstrated how AI can dramatically reduce information processing time, allowing clinicians to move from hours of information gathering to immediate clinical decision-making. This represents not just efficiency gains but fundamental workflow transformation that could improve patient outcomes.

The discussion also explored how AI could enable healthcare workforce optimisation by allowing general practitioners and frontline workers to handle cases that would traditionally require specialist referral. This capability could significantly improve healthcare access in regions with limited specialist availability while reducing system costs and patient wait times.

Data Governance and Trust Building

A critical theme throughout the discussion was the importance of building public trust in AI healthcare systems. Dr. Aditya Yad emphasised that data trust and privacy concerns remain ongoing debates that must be resolved before widespread AI adoption can occur. People need confidence in how their medical data is collected, stored, used, and protected within AI systems.

This trust-building challenge extends beyond technical data security to encompass broader questions about AI governance and accountability. Healthcare systems must demonstrate not only that AI tools are technically safe and effective but also that they operate within ethical frameworks that respect patient autonomy and privacy rights.

The panellists recognised that different regions may have varying approaches to data governance, but establishing trust remains universally critical. Success in building this trust could determine whether AI healthcare solutions achieve widespread adoption or remain limited to early adopters and specific use cases.

Future Outlook and Strategic Implications

Looking towards the next five years, the panellists expressed optimism about AI’s potential to transform healthcare while acknowledging significant implementation challenges. Chris Ciauri noted that Anthropic releases new, more capable models every 2.5 months, with each iteration being exponentially more intelligent and powerful than the previous version, including the recent Claude 4.6. This rapid advancement suggests that current AI capabilities represent just the beginning of what may be possible.

The discussion revealed an emerging consensus that successful AI healthcare implementation requires coordinated efforts across multiple dimensions: technology development, policy frameworks, workforce training, and economic incentive alignment. The recognition of countries like India as innovation leaders rather than just markets represents a significant shift in global technology discourse, suggesting that solutions developed for emerging market challenges could ultimately benefit healthcare systems worldwide.

Conclusion

This comprehensive discussion demonstrated that AI’s transformation of healthcare is not a distant possibility but a current reality requiring immediate attention to implementation challenges. The conversation successfully moved beyond typical technology hype to address practical concerns about safety, workforce development, economic sustainability, and global equity in healthcare access.

The panellists’ diverse perspectives—combining clinical experience, technology expertise, and policy insight—created a nuanced understanding of both opportunities and challenges. Their collaborative approach and focus on practical implementation over the next 3-5 years suggests a mature and responsible approach to AI healthcare development.

Perhaps most significantly, the discussion reframed the global healthcare AI narrative from one of Western innovation being deployed to emerging markets to one of collaborative development where countries like India serve as leaders and innovators. This shift towards more inclusive and globally representative AI development could ultimately benefit healthcare systems worldwide, ensuring that AI solutions address the full spectrum of global healthcare challenges rather than just those prevalent in developed markets.

The path forward requires continued collaboration between technologists, clinicians, and policymakers, with safety and human-centred design remaining paramount. As the discussion concluded with an invitation to continue these conversations at next year’s AI Summit in Geneva, Switzerland, it became clear that the global healthcare AI community is committed to thoughtful implementation that maintains trust while ensuring equitable access to AI’s transformative potential for global health outcomes.

Session transcript

Dr. Sabine Kapasi

The last day of the event is a little slow today. You know the energy of the last three days seems to have gotten people a little right. A big week. Yeah, I know. So today, unfortunately, we don’t have a couple of people who are supposed to be here, namely Rizwan sir as well as R.S. Sharma sir. Both of them stalwarts in the industry of setting context in both Indian healthcare systems but also in setting up global standards for digital public infrastructure in India. But let’s make do without them for today. So today, we are talking about AI in healthcare, right? I’m Dr. Sabine Kapasi. We have Dr. Aditya as well as Chris here with us.

I’ll give their intro in a bit. We recognize today that AI will transform healthcare. Given that India and many other… Other low – and medium -income countries have very low levels of digital adoption, though… it’s important to determine where AI solutions are likely to have the largest ROI rather largest opportunity in the next 3 to 5 years. So in addition we also need to ensure that doctors, hospitals and other healthcare professionals are getting ready to leverage AI as well. So today we are going to focus on identifying near term opportunities for India and India as a leader in the LMIC space. That’s low and medium income countries space. And discuss strategies to strengthen the healthcare system for adoption of real use cases of AI.

I think that’s going to be one of the challenges as well as one of the longest value gain that we are able to deliver as we go. So before we go ahead I would love to introduce my co -panelists here. Chris is the managing director at Anthropic. He leads. Global expansion across EMEA, APAC and Latin. with over 25 years of experience scaling SaaS and cloud businesses, including senior leadership roles at Salesforce, Google Cloud, and most recently as the CEO of Unilever, he brings deep expertise in enterprise AI adoption and national technology growth. He is known for building high -performance global teams and driving transformation through collaborative leadership. Thanks a lot, Chris, for being here.

Chris Ciauri

Thank you for having me.

Dr. Sabine Kapasi

Before we introduce Dr. Aditya, I would love to throw a question to you. So how does Anthropic, as a company now, view opportunities in healthcare AI, not just in the U.S. and in Western Europe, but also countries like India and the global south, of how AI is being adopted, especially in the healthcare industry?

Chris Ciauri

Thank you for having me. I’d say we think healthcare, is certainly one of the areas where we’re going to be able to do a lot of things. AI can do a lot of good. It also can create a lot of harm if done carelessly. And Anthropic was founded with a mission around safety, and we focus a lot on that. So we like the tension between capability of AI models but also making sure that the safety is right so that we can deliver on some of the opportunities. I think maybe I’ll use two examples just to frame areas that we think big impact can happen. And I’ll use a U.S. example, I’ll use an India example.

If you think about certainly one of the biggest challenges in the U.S., India has this too, Sangeeta mentioned some of this, but it’s really around the burden of administration. So in the U.S., only 30 % of a clinician’s time, a doctor’s time, is spent on patient care. The rest is on paperwork and administrative tasks. I think in India, one of the biggest challenges is just access. So, you know, there’s data over the last decade that says that, you know, the average primary care visit only lasts two minutes. So if you think about where AI can impact those, and we believe it can have a huge impact, you know, if we can decrease the paperwork, decrease the administrative burden, we can have doctors in the U.S. and other places spending much more time on patient care.

Huge outcome. Can have phenomenal ROI. In India, you know, we think we can help. Solutions like ours can help. Make your health. Care system much more broadly accessible. And the other thing that’s uniquely exciting about India is. you’ve built a digital healthcare system that’s the envy of the world. And we look at that with excitement because we think that gives the AI a really great place to land when you’ve got that kind of digital infrastructure. So maybe my last comment for those that sort of don’t know Anthropic and don’t know we’re up to, we’re so excited about opportunities like this that we announced that we launched our operations recently. Recently, we’ve opened an office in Bengaluru because we think to address a problem like this, we want to be here on the ground building with you.

And we also think, you know, people have talked about the scale of India, leader of the global south. If we can make this work in India, we think we have the possibility to shape how AI -driven healthcare evolves in the rest of the world.

Dr. Sabine Kapasi

No, I think you’re right. Right, and you have worked with every tech company under the sun. which is amazing. Someday you’ll have to tell me what that looks like because, God, I’m a little further away from tech. I’m a clinician by…

Chris Ciauri

I’m as far away from being a clinician as you are from being a technologist.

Dr. Sabine Kapasi

I think that shouldn’t be so, right? And coming back, that is where I think it would be great to introduce Dr. Aditya Yad. So Dr. Aditya is the India Relations Advisor at Invalude, the innovation and investment promotion agency of Canton Broad, Switzerland. Based in Lausanne, he plays a strategic role in strengthening Switzerland -India collaboration by facilitating cross -border partnerships, supporting high -growth startups, and enabling market entry for Indian companies into the Swiss and European innovation ecosystem. He himself is a biotechnologist and has worked in the interaction or rather at the cross -section of tech, investment, and innovation. He has done a lot of investments and I believe biotech as well. With a focus on technology and research -driven enterprises and global expansion pathways, Aditya acts as a key bridge between Indian entrepreneurs, investors, academic institutions, and the vibrant innovation landscape of what is more as vibrant?

Honestly, I doubt it. India is far more vibrant, to say the least. We can have a debate on that, yeah? Yeah, no, maybe we’ll discuss that in a bit. But, you know, as we mentioned, you are as far from being a clinician as I am probably from being a technologist, but we need a middle bridge. And when we are talking about healthcare and AI systems, we need a middle bridge. So you have looked at ecosystems on both fronts, right, and innovation happening on both fronts in India and Switzerland. Switzerland having a deep research in biotech as well as a deep legacy of research in biotech, and now adoption of new technologies. New technologies on top of that legacy versus India who has leapfrogged into an area of… of fast growth and fast technology adoption.

How do you see these two systems playing out and interacting with each other for a larger good in outcomes, especially when we are looking at health care?

Dr. Aditya Yad

Thank you for the question and for the invitation. So, you know, as you said, Switzerland and India, when you look at the size of the country, the size of the population, of course there will be very different challenges for both countries. On the Swiss side, to continue on the debate, you know, Switzerland has been ranked number one in the Global Innovation Index for the past 15 years straight. And largely part of that is thanks to the health care industry, the biotech, the pharma, the life census industry. Today we have around 1 ,700 companies or research institutions based in Switzerland for a small country like this that is really giving this vibrant ecosystem of innovation that we have.

The second point also is that, you know, Switzerland is not a big domestic market, right? We are 9 million people. So all the products…

Dr. Sabine Kapasi

That’s not even Delhi. Like, that’s not even Delhi.

Dr. Aditya Yad

That’s why I usually like to have this scale, you know. Somehow, you know, by just a parenthesis, so India and Switzerland have signed this free trade agreement now, right? So we are concretely in business between Switzerland and India. Part of that free trade agreement that was signed by both governments, now Switzerland and the EFTA countries now has a commitment to invest $100 billion into India in various sectors. That includes also healthcare, by the way. And to create also 1 million jobs, direct jobs in India in the next 15 years. So now there’s a concrete engagement with both countries. So when it comes to healthcare, you know, Switzerland is known for two things. A very efficient and highly qualitative healthcare system, but a very expensive healthcare system.

So you get the price and the quality that goes with it. So this is where AI is actually going to play a very big role. If you talk about cost reduction, optimization of all the processes from research to putting medication on the market, using technology, using AI will tremendously, we believe, bring down the cost of health care. People are now, because I’m also a policymaker in Switzerland, I’m a legislator, in the public debate there’s a lot of heat or a lot of pressure from the public that health care premiums are too high for what they’re getting. So this is exactly the very easy definition where we can say, okay, now we have these tools that can accelerate drug development, not to spend a few billions in developing a new drug, but using AI tools to speed up the process, to increase the probability of finding the right targets and clinical validation and market access.

So this is where I think there is a very big potential. And from the industry’s perspective, we also see that a lot of companies are now So either shifting from traditional pathways into AI initiatives, or the smaller companies now, the startup companies with which we work, they include and they embed AI strategy within the development of their company overall. So it’s become completely normal that AI has to be included from the very start. And this is also what startup companies, new innovative products, they’re using in their own pitch in order to convince also investors to get these investments. Last year, I just published this report. So we had $2 .5 billion of investment going into Swiss startups just for last year.

Many of them are using and have been able to raise funds because they’ve been integrating AI tools in their development in that sense.

Dr. Sabine Kapasi

Thank you so much, Aditya. Chris, back to you. So first of all, I’m really glad that we have people who have been making for healthcare systems but are not native to healthcare. Because… sometimes when we think about healthcare, we only think about doctors, right? Or we think about hospitals. But thankfully, we know that healthcare is so much more. It’s not just about doctors or hospitals. And you as a company, of course, as I said, you have worked across, so please feel free to share your experience across several different domains that you have worked around, but have worked with several technology companies that were not native to healthcare, but now see healthcare as a huge opportunity as well.

So which are the specific use cases that companies like Anthropic view or are targeting for to solve healthcare problems that they are looking to solve? And how do you test for the risks, especially when you’re building LLMs which are quite generalized? Because healthcare has a very immediate outcome of risk, and that’s something that needs to be tested for or at least covered for. So how do you guys look at it?

Chris Ciauri

Maybe I’ll do the risk first, and then I’ll talk about a few use cases. And by the way, thank you for the comments that basically date me, because you know all the companies I’ve been around a long time. But I think I’m privileged to be part of Anthropic and what’s going on right now in AI, because I think by far this has the opportunity to transform health care more than any other technology transformation that we’ve seen over the last three decades or so. But coming back to the risk point, I think I made the point up front that AI can do a lot of good in health care, and it can do a lot of harm if you’re not very careful about the way you use it.

I think because we’ve been so focused on safety, Claude, uses language like, I don’t know. and I’m not certain quite freely. And we think that’s critical in an industry where the stakes are so high. And I’ll give you one example. One of our customers is Banner Health in the United States. They’ve used us to summarize 100 -page oncology reports where previously a clinician comes in, and they’re getting information that was across multiple appointments and specialists, and it took them eight hours just to get to the point that they could start to provide an opinion, care, judgment. That is now summarized concisely. So all of that time, all that administrative time or information retrieval time, is now quickly moving into judgment and delivering care for patients.

So that’s both a use case, but it’s also like, you know, I think a demonstration. of why did they choose us? Ultimately, the final decision came down to they wanted a model that was so based in safety that it said, I don’t know, or I’m not certain. What they didn’t want was a model that was confident or felt confident, but it was wrong. So I think safety is paramount for us. We think it’s table stakes in health care in everything we do. Maybe use cases, and I briefly hit on this before, but I think certainly administrative burden is one, and we think that’s pervasive everywhere. That speaks to your take costs out of the system issue.

The 70 % of time that I talked about doctors in the U.S. spending on admin, that’s a $1 trillion problem or opportunity. So the magnitude, if we could start to address, things like that with AI globally is huge. A second one, drug discovery. So in some of our work with customers like Novo Nordisk and Sanofi, we’ve been able to reduce drug development, life cycle times, heavy paperwork, heavy regulation from eight weeks to eight hours. Like just phenomenal difference in terms of how quickly we could get amazing drugs and medicines to market if that becomes pervasive. The last one, I’ll be really India -specific. You know, I mentioned access here is a challenge, and that certainly if the country can get to the point where they can get the health care system to serve all Indians, that would be game -changing for India and I think for the global south.

One of the big barriers is multilingual. So. So you can’t use a model that’s good in English, but it’s not good in other languages. So as part of our entry into Indian over the last six months, we’ve trained Claude on 12 Indic languages, and that’s not to say that it’s done and over, there’s more languages, there’s dialects, but I think those are the types of things where AI can improve access for health care.

Dr. Sabine Kapasi

No, I think just about 15 days ago when you guys launched your Pro model, Pro Max, I think that was the one, right?

Chris Ciauri

It was a couple of weeks ago, it was our version, Claude 4 .6.

Dr. Sabine Kapasi

Yes, and I had a friend call me up and seeing the stock market very heavily when you guys launched your recent version, and they’re like, you know, there are jobs that are in serious danger. So you should actually, and I remember… I remember telling my team to go back and start playing around with… with the new version because the 11 attachments that have come through have been fascinatingly interesting in the way they are going to be adopting new workflows. But one of the things that EI is adding a lot of value in is in B2B space. So before we go back to that and we’ll talk a little bit more about how EI is shifting biotech as an industry because healthcare is not just about patient delivery but it’s also about how we get there as you touched upon in drug discovery.

And I think AlphaFold was a phenomenal change the way new protein discoveries are being done today. And I mean I could not believe that in my lifetime I have seen such a jump in technology and there is so much more to come. So it could not have been possible without large scale models. But that being said, when we look at countries like India and this is when I was practicing as a surgeon, I used to see my OPD used to consist of 200 people a day. So the amount of time we spend per patient in the lower half of the world, let’s put it that way, is very different. The workflows are very different. And even though the clinical logic might remain the same, the clinical skills deployed are extremely different in terms of action items.

So, once we, I’ll circle back to you on that, but we would love to discuss how you are adopting those kind of use cases to deliver value, not just in countries which have optimized for outcome, especially optimized time for outcome, but use the same principles for a global scale outcome difference as well. But before we do that, I would love to have your thoughts about how, in your perspective, you know, drug discovery, clinical trials, regulatory sciences, as well as manufacturing, is being affected by the new innovations in AI, and how do you see that happening in the next 10 years? Or five years, I think. AI is moving fast enough. We can’t predict 10 years at this point.

Dr. Aditya Yad

Hello, hello. One thing, so we touched upon drug discovery and the impact of AI in the healthcare system in general, so with hospitals, with clinics, and so on. Manufacturing is actually very interesting, and it’s also very relevant to India because there’s a new policy also in India to have this biofoundry. So biomanufacturing in general in India has become a national priority. And the inclusion of AI tools into this is very, very relevant. We see some companies already shifting into that also in Switzerland, also smaller companies. If you really talk about the manufacturing of different products in a controlled environment where parameters are being monitored over time and self -learning about optimizing those parameters in order to increase yields, for instance, or increase.

The ROIs on these production systems is becoming a very interesting trend. We have a bunch of companies. We have the large companies, Novartis, Roche, Lonza. producing massively but you now have these what you call bioreactors that are smaller in size but more qualitative because they can use the AI tools and so you are able to produce very highly qualitative products that could be very expensive but because AI is being used in a small control with a better yield the prices and the production costs are being limited so this is typically one example where from the very industry’s perspective there is a big interest and a big potential over the next few years because at some point these drugs they will have to be manufactured, they will have to be distributed they will have to reach the patients at the end of the day this is the main goal right so how do we streamline all the process from R &D to the drug delivery at the patient level all of these can be used and AI being infused into all of these different steps that will be the challenge but on manufacturing I think there is a lot of potential in the next five years.

Dr. Sabine Kapasi

Thank you so much. I think one of the things that also has a massive potential is screening and also health from an insurance and finance perspective. One of the things that are shifting a lot of use cases and creating a lot of different frameworks are the way we now have diagnostic technology evolve so fast that we can take it to people’s homes directly and make sure that the signals that we are capturing or biomarkers that we are discovering at pace reduces costs of testing drastically so that screening as a solution becomes available across the world and not just specified to areas where large capex into diagnostic capacity building is being evolved. So I think that is one thing I would love to have some thoughts on.

I’m so sorry, I was off there for a bit. Okay. As we spoke about in a bit, that, you know, you said countries like India and other low -income countries, or rather other global south countries. I mean, I would never say low -income anymore. Look at us right now. But other global south countries are shifting now in terms of adoption. And at least countries like India have massive digital adoption coming through as well. In urban areas, there’s close to 90 % smartphone ownership. In rural also, it’s touching 75%, which is just fabulous and mind -boggling altogether. But in such ecosystems, how do you see the potential of such markets actually developing solutions within healthcare that shift the perspectives of the low -wage, or rather in the developed markets as well?

Chris Ciauri

You know, this, I’ll share some statistics that you might find surprising based on what you just said. So India has the highest adoption of cloned. outside the U.S. It’s second in the world for cloud adoption. So, yes, it’s global south, but your country is consuming AI. It’s probably all of my travels I’ve not seen or felt a country that’s more optimistic about the potential of this technology. It’s also, by the way, it’s second in usage, but it’s the fastest growing. So in the last four months, the usage of cloud and the revenue for Anthropic has doubled in India. So I wanted to make that point. But I think if we think about what does that mean, I think that means given the context of an amazing digital health record system that you’ve built, which is not just unprecedented in the global south.

It’s one of the few globally. And it really does give you and us, I think, the ability to… to do something quite special across… you know the largest democracy in the world a massive population that’s also uh got the additional challenge of multilingual and and it’s why i said at the beginning i think if if you can do something you know we can make this work in a place like india not only does this sort of give the the global south i think a model i think it gives the whole world a model of how we could really see health

Dr. Sabine Kapasi

care transform exactly i think some of the ways for example if you um i’ll take a you know take a case from my own stories um we had a patient who was a dengue patient we knew clinically there’s a dengue patient we were working in a very very remote region so we had no access to diagnostic tests at that time and we just started treating them the reason is the drugs were cheaper than the diagnostic tests and the patient could afford any so for the system at that time it was a trial and error problem but the clinical values were stark enough for us to know that this was not actually a risky case. This was almost certainly a dangerous case, even though we didn’t have the data to back it up.

So I think taking those clinical validations and those clinical intelligence, combining them with the new discoveries and biomarkers that now with new technologies are coming in every day and making diagnosis and drugs affordable across the world, that I think is going to be the next big leap in the healthcare ecosystem, and it is going to make a fundamental shift within that ecosystem, at least in my understanding. And you guys are at the front

Chris Ciauri

I completely agree. If you think about that situation you described there and the one before when you said you used to see 200 patients a day, I’ve been very lucky. I think those of us that work in technology companies that get to help businesses and different sectors. do things better. I always feel like we’re lucky because normally I’d get to spend 30 minutes with you and I’d say, tell me what’s going on. Like if you could, if you could fix one thing, what would, what would mean that, um, you know, you’d be so efficient in the care that you gave that you could extend the time with patients and you could provide a way that a hundred of them could, could self -serve.

Um, I mean, I think those are the kinds of conversations that we get to have, um, as this technology, uh, you know, hits a sector like healthcare.

Dr. Sabine Kapasi

No, of course. Uh, and he’ll take it off the stage later, but, uh, to your note, you have been a biotechnologist and a researcher yourself. How do you see, and we all know in healthcare prevention is better than cure. I mean, that’s been said, but as a system, we have never really adopted it at scale. So how do you see diagnostic technology evolving through the use of AI in the next few years? And how do you see that playing out for countries

Dr. Aditya Yad

diagnosed and prevented from the treatment. So the cost of having maybe a diagnostic tool that costs two or three or four times more than the current one was not enough to convince insurance companies to pay for that in order to avoid a $20 ,000 or $50 ,000 therapy for cancer for instance. So the narrative and the whole system is being built around that, around treatment. If we can use and prove that AI can have a significant impact both on the quality of the diagnostic but also on the cost of developing these diagnostics and being put on the market, then I think we can really make a change in terms of how we see healthcare as a system as well.

Dr. Sabine Kapasi

No, I think that’s true. So we have been talking to medical device companies who are now targeting new age diagnostic tools and even companies, legacy companies like GE and Philips. And one of the larger problems that they face is that adoption of diagnostics is an issue because you essentially are talking about, especially in screening and very much so in screening, because if you feel there is a problem with your body, you will certainly want to figure out a way to solve it, be it through self -pay models, which is prevalent in this part of the world, or through insurance models. But screening basically precedes the need for healthcare. It precedes when you feel you have a problem with your body.

So how do you make people pay when they don’t see the need for it? So I think one of the things, throwing that back to you, is two things, two questions. One point is that, you know, healthcare is an industry because we are quite, we understand the risks of using any tool, be it pharma tool, be it AI tool, in the healthcare industry, we know the risks are quite immediate and sometimes life -threatening. So we are always quite skeptical in how we position ourselves. We position the tool. So how do you train the healthcare workforce to adopt it while also keeping a layer of difference within the tool as well as people adopting it? Because you don’t want people acting on health advisors from an LLM today.

No, of course. That’s one. So how do you create the education for the healthcare workforce strong enough, but also ensure that people are not directly acting on those advisors? And secondly, how do you educate the ecosystem better so that such processes like screening and making sure that you precede the need, especially in healthcare, precede the felt need in healthcare? How do you rather execute those kind of strategies? These two are questions I would love for you to give me a thought on.

Chris Ciauri

Maybe on the second one, I might see if you have an opinion because I’m an AI person and not on the clinician side. On the first one, I think. I think you have to be clear, and we certainly have clarity on this in dealing with health care systems and customers in the health sector and pharma sector. You know, AI is for preparation. Clinicians are for judgment. So we have no intention of Claude being a doctor or a nurse. Thank God. Doctors are really scared. I think you have to be clear about that. And it goes back to what I said before, because these models, if they’re really going to serve health care, you know, Claude’s going to know when to say, I don’t know.

I’m not certain of that. Like, this is where you need to go, you know, and we’re going to have this conversation together with the clinician. And we think that’s table stakes. The same thing with, you know, Claude will never use someone’s patient data to train our models. And I think that’s the key. And I think that’s the key. If you don’t make those your non -negotiables, then AI will not get to have the impact that it should have.

Dr. Sabine Kapasi

No, that’s true. And if you can answer that.

Dr. Aditya Yad

Maybe just to weigh in on that, because it is true what you just said. So the importance of AI being with the companies, with the innovation segment, and not necessarily at the end. Because, as you said, there can be also pros and cons on that usage. So, for instance, I’ll give you one practical example with our state government. We launched a program just one year ago because we noticed that especially smaller companies, small mid -sized companies developing these new technologies, how did they embrace AI? And there is still a little bit of caution about how to use that, how to not make a mistake from the beginning of the process and then go into a direction that was not anticipated.

So we launched a program a year ago where we have cohorts of companies that is being used. We have a lot of companies that are being funded by the state government. where we talk directly with the CEOs of these companies, and we have a leadership program to train CEOs to think how they can implement AI from the very start of this process. The challenge we have is that in our state, for instance, we have 40 ,000 companies, SMEs. How do we convince them to go and to embrace AI, and how do we sell them the benefit of using AI, because there’s also resistance to change on that level. So as long as we don’t know who is going to really take the lead on using the AI tools, then everybody will be using more or less of it, and then there will not be a homogeneous application of AI.

So here we said, okay, the industry needs to take the charge of doing that, and then we’re going to train the people, as you said, on the use of AI, not just as a technological tool, but as a strategic roadmap for the company going forward.

Dr. Sabine Kapasi

So, see, we are in 2026 today. four years away from 2030 and a lot of plans that a lot of countries have made for adoption in 2030, so India included. There are a couple of areas like workforce enablement for AI adoption, especially when we are talking about healthcare. We are looking at solving B2B cases, workflow management, time enhancement of the skilled workforces and also create some level of skilled enhancement of creating bots or other agentic systems which can in some proportion aid people who are not as credentialed, let’s put it that way, so enabling a frontline workforce or enabling a GP to solve some cases which may otherwise be referred to a higher center. So reduce burden and distribute the burdens in healthcare systems.

I think that is one use. In this case, that is something that I would love for you to throw some light on, Chris. and I think also the non -sexy use cases, like the drug discovery cases, which are phenomenal not just for business but also for changing the world as we know it, and diagnostics. So if you can throw light on these three verticals and how you see them panning out in the next five years, that would be really, really helpful.

Chris Ciauri

I mean, I might sort of frame it with this. I think, and you talked a little bit about it. Your friends have kind of seen what Claude, sort of the latest version of Claude, and what we’ve seen in the last, Anthropic’s a five -year -old company. We’ve had a commercial model, or a frontier lab. We’ve had a commercial model on the market for coming up on three years, and each model, which now we’re at a rate of every two and a half months a model releases, is exponentially more intelligent, more powerful than the last one, and safe. And I think that’s really important. we don’t see that stopping. So I think what makes me extremely optimistic about the ability to really transform health care on many dimensions that you talked about is this technology will get better and more enabling.

As long as we do it incredibly safely, the benefits are probably hard for us to imagine with how fast it’s moving, but I think it’s a tremendous opportunity.

Dr. Sabine Kapasi

Just before we close this, one more thing. LLMs versus small language models, right? Targeted use cases. How do you see the industry evolving in the next five years in health care? Targeted use cases?

Chris Ciauri

Yeah, I think what you will likely see is smaller targeted use cases will have a place. You know, maybe that’s out on the edge in specific things, and open source will have a place in that. I think as a frontier lab, we have one model. It’s Claude. It comes in a few versions, you know, so that it can be scaled up and down for different use cases. Our position in the market is, like, let’s make Claude the most capable and safe model that we possibly can. Let’s keep that exponential innovation going, because our place in the market is going to be to drive the greatest amount of innovation and transformation. And there will definitely be a place for more edge use cases with smaller models.

I just think it’ll be a great place. It’ll play out differently.

Dr. Sabine Kapasi

And you see countries like India playing out an interesting role in that? Sorry? How do you see countries like India playing out an interesting role in that?

Chris Ciauri

Yeah, I mean, I think we’ll see many countries probably playing more on the small language edge use case side. Today, the frontier is in a couple of countries, but I think there will be opportunities that we can’t see.

Dr. Sabine Kapasi

Thank you so much. And Aditya, as a policymaker, if you can throw a very quick light on where do you see the next five years spanning out in terms of, and what are the things that we will need to be careful about that we don’t see today?

Dr. Aditya Yad

I think in general, the trust around data and personal data, medical data is still a debate. So this is ongoing. There’s a lot of awareness building to be made. We have to gain the trust of the people that they trust the systems, what is happening with the data, where is the data flowing, and how do they see the ultimate benefit from them. using this data. From that point onwards, we can really build different systems, we can think about new things, but that is still something that we have to work on.

Dr. Sabine Kapasi

Thank you so much, Aditya. Thanks, Chris, for joining us, and essentially creating equity and AI, which is useful for all, especially for something like healthcare, is something that we all strive for and hope that this is going to change the world in the next five years. Thank you so much for joining us for this chat. Namaste. Thank you. And see you all in Geneva next year, because the AI Summit will be in Switzerland next year. So we’re all welcome there as well. Yes, of course. Thank you. Thank you. Thank you.

Chris Ciauri

Speech speed

148 words per minute

Speech length

Future Outlook: Model Evolution and India’s Strategic Role

Topics

Artificial intelligence | Closing all digital divides | The enabling environment for digital development

Dr. Sabine Kapasi

Speech speed

152 words per minute

Speech length

Future Outlook: Model Evolution and India’s Strategic Role

Topics

Artificial intelligence | Closing all digital divides | The enabling environment for digital development

Dr. Aditya Yad

Speech speed

174 words per minute

Speech length

Future Outlook: Model Evolution and India’s Strategic Role

Topics

Artificial intelligence | Closing all digital divides | The enabling environment for digital development

Agreements

Social and economic development | Financial mechanisms

Data governance | Human rights and the ethical dimensions of the information society | Capacity development

Unexpected consensus

Arguments

AI Model Evolution

Summary

Chris advocates for frontier labs focusing on large, general-purpose models like Claude that are continuously improved every 2.5 months, while acknowledging that smaller, targeted models will serve edge cases. However, there’s an implicit tension about whether the future lies in large general models or specialized smaller models for healthcare applications.

Topics

Artificial intelligence | The enabling environment for digital development

Implementation strategy for AI adoption

Speakers

– Chris Ciauri
– Dr. Aditya Yad

Arguments

Human-AI Collaboration

Strategic AI Implementation

Summary

Chris emphasizes clear boundaries where AI handles preparation and clinicians handle judgment, while Dr. Aditya focuses on training CEOs and leadership for strategic AI implementation from the start. These represent different approaches to AI adoption – one focused on maintaining professional boundaries, the other on comprehensive organizational transformation.

Topics

Artificial intelligence | Capacity development | The enabling environment for digital development

Unexpected differences

Data governance priorities

Speakers

– Chris Ciauri
– Dr. Aditya Yad

Arguments

AI Safety Priority

Data Privacy Trust

Explanation

While both speakers acknowledge data and safety concerns, they emphasize different aspects. Chris focuses on AI model safety and uncertainty acknowledgment, while Dr. Aditya emphasizes public trust in data governance. This represents different priorities in addressing AI safety – technical model behavior versus public acceptance and trust.

Topics

Data governance | Human rights and the ethical dimensions of the information society | Building confidence and security in the use of ICTs

Overall assessment

Summary

The discussion shows remarkably high consensus among speakers on the potential of AI in healthcare, with disagreements mainly centered on implementation approaches rather than fundamental goals. The main areas of difference involve technical strategies (large vs. small models), organizational implementation approaches (boundary-setting vs. comprehensive transformation), and priorities in addressing safety concerns (technical safety vs. public trust).

Disagreement level

Topics

Social and economic development | Financial mechanisms

Data governance | Human rights and the ethical dimensions of the information society | Capacity development

Takeaways

Key takeaways

AI has transformative potential in healthcare through administrative burden reduction, improved access, and accelerated drug discovery, but safety and human oversight remain paramount

India’s digital healthcare infrastructure and high AI adoption rates position it as a leader for global healthcare AI implementation, with potential to influence worldwide standards

The Switzerland-India partnership, including a $100 billion investment commitment, creates significant opportunities for cross-border healthcare innovation collaboration

AI applications span from reducing drug development timelines from weeks to hours, to enabling multilingual healthcare access, to optimizing manufacturing processes

Healthcare workforce education must balance AI adoption with appropriate clinical skepticism, maintaining the principle that AI supports preparation while clinicians retain judgment authority

Prevention-focused healthcare and screening programs face economic challenges that AI cost reduction could help address, potentially shifting from treatment-centered to prevention-centered systems

Resolutions and action items

Anthropic has opened operations in Bengaluru and trained Claude on 12 Indic languages to address multilingual healthcare access challenges

Switzerland launched a CEO leadership program to train companies on strategic AI implementation from the start of their development process

The discussion established that AI models in healthcare must acknowledge uncertainty and never use patient data for training purposes as non-negotiable safety standards

Unresolved issues

How to effectively educate healthcare ecosystems to adopt screening and preventive care when patients don’t feel immediate need for healthcare services

How to scale AI training and adoption across thousands of small and medium enterprises in healthcare (Switzerland faces challenge of reaching 40,000 SMEs)

Data trust and privacy concerns remain ongoing debates that need resolution before widespread AI healthcare adoption

The balance between frontier AI models versus smaller, targeted language models for specific healthcare use cases needs further development

How to ensure homogeneous and strategic AI application across healthcare systems when adoption rates vary significantly

Suggested compromises

AI should be positioned for preparation and support while clinicians maintain final judgment and decision-making authority

A hybrid approach where frontier AI models handle complex cases while smaller models serve edge cases and specific applications

This challenged the framing of India as a ‘low-income’ country and revealed it as a leading AI adopter, contradicting assumptions about technology adoption patterns in the Global South.

Impact

This reframed India’s position from a recipient of technology to a leader in adoption and potentially innovation. It led to discussions about India as a model for other countries and the potential for reverse innovation flowing from Global South to developed markets.

Overall assessment

Explanation

This addresses healthcare workforce optimization and access improvement by using AI to enhance capabilities of less specialized healthcare providers

Disclaimer: This is not an official session record. DiploAI generates these resources from audiovisual recordings, and they are presented as-is, including potential errors. Due to logistical challenges, such as discrepancies in audio/video or transcripts, names may be misspelled. We strive for accuracy to the best of our ability.