A free primary care physician for every Indian available 24×7 and a free tutor for every child is doable in 2-to-5 years: Vinod Khosla | India News

Between the perils of tomorrow and promise of tomorrow, Vinod Khosla has always bet on the latter. That’s why billionaire founder of Khosla Ventures, who describes himself as a ‘techno-optimist with caring & empathy’, is busy convincing the world of an AI-powered future that can be good for all. Among the first investors in Open AI–and several Indian AI startups–Khosla spoke to Rohit Saran, Sujit John and Surojit Gupta of TOI on a host of issues ranging from AI to startups to Mohammand Yunus and Elon Musk.Here’s an edited excerpt:
You have been one of the early investors in OpenAI. You have also invested in AI startups in India. As a longtime technologist and tech investor what are your views on AI and the opportunities it offers?
Let’s look at a couple of different aspects of AI. From the national security point of view, governments want to be in control of their AI destiny, not be dependent on foreign algorithms. This is why we invested in Sarvam in India. Because India should have a national AI champion. We have invested in a company called Sakana in Japan.
Japan has already announced doubling its national defence budget. They made it clear that AI be a Japanese AI, so they have control over it. Sovereignty and AI are important at a level people don’t quite understand.
I would like to see a primary care physician for every Indian, available 24/7, for free as part of the set of digital services the government offers. Aadhaar was the first example of such a service. The hardest part of that service was identity, and that’s already done. With UPI payments are done too. There’s no reason not to add other digital services — the two most important ones are a free tutor for every child and a free primary care doctor for every citizen of the country.
I would like to send a clear message to the Indian government that AI educators, AI teachers for every child in India would cost far, far less than today’s schools and provide much more personalised instruction. Similarly, AI primary care doctors could scale primary care and health far better than any other way.
None of these are hard technologies in today’s context of AI, and we are working on both. If the Indian government wanted to take and offer it as their service, we would give them all the technology for free.
There’s no agenda here. I’d give them all the technology for free, so it becomes a government service, and nobody has any questions about what the agenda is, other than these services should be available to every Indian and every Indian child and even adult education. It’s very, very doable in the very short term, and I mean two to five years max.
Where does one start? What would be the roles for the government, private sector?

My wife is in India right now, talking about our free AI tutor (CK-12 tutor). One way is to just take it over. It’s used in many countries already. It’s being translated into 120 different languages. I’m fine if they develop it inside India and take it over and have a starting point that’s useful today.
Kids using this tutor do better. Turns out it’s not only cheaper, but also substantially better. It can teach each kid at the level at which they are, in their language, in a way that the kid will understand best. For example, the same physics concept can be explained using the analogy of cricket or dance depending on which kid understands better with what analogy. You can generate AI explanations in real time, generate the content so it looks more interesting, though it’s teaching the same concept.

An AI tutor would change the level and method of teaching depending on each student’s level of understanding. If they’re not keeping up with the class. Teach them in simpler steps, change the language…
I really hope this becomes a national agenda in India. We would give the technology. The content is already there. I think fairly large chunks are translated into the Indian context. It’s relatively lightweight to launch it as part of the digital public infrastructure.
Digital public infrastructure should include payments, which are basic financial services everybody needs. It should include education. It should include primary healthcare. You can then expand it to others. This can be done, and it can be done philanthropically.
Medical diagnostics in rural areas is an issue, not so much getting a doctor online. How do you think AI can solve this problem?

In India we started ‘10 Bed ICU’ during Covid, because ICUs were not available in most parts of rural India. I donated the first couple of 100 to remote areas and we wanted to cover all 700 remote districts in India. It’s working really, really well.
AI lets you take an unskilled person and teach them very quickly how to do skilled tasks, because the expertise sits in the AI. A 10-bed ICU is a very simple version of it, typically under the supervision of a regional hospital. A hospital in Manipur would cover the remote areas around it. It’s just the beginning of what is really possible.

Yes, if you need an MRI, you are going to go to a city to get an MRI. But for simpler things like even an ultrasound can be done by an AI. You don’t need an ultrasound technician anymore. You can buy these devices for Rs 50,000 rupees. Of course, there’s expensive ones too, but small handheld, portable ones are becoming inexpensive and adapted to the Indian environment. All 6 lakh, however many villages there are, should have one. In fact, these diagnostics can be put in one van — blood test, ultrasound, ECG…
We have a company called Alivecor that sells ECG machines in the US for $99. They diagnose without a cardiologist 35 different cardiology conditions with AI. It’s FDA approved. Most of cardiology can be done with a simple device, and you don’t need expertise, and expertise is what’s expensive. The cardiologist is what takes 20 years to train. And so this is the promise that I hope India pivots to rapidly.

India has large youth unemployment and the fear is that AI could worsen it…

This is a real issue. Firstly, you can upgrade the skills of India’s youth very quickly with AI. We had a company (that General Electric bought) to teach an Uber driver to do a cardiac ultrasound in a week. It’s FDA approved. This is an extension of the idea of an AI tutor, which is where we should start. You can impart skills, which very quickly, can be upgraded so they can participate at a higher level. The countries that do this well will be more competitive than others.
Secondly, it’s hard to imagine the kind of abundance we can get (in an AI-driven economy). If GDP grows 10% in real terms for the next 25 years, you can imagine the standard living we can achieve and the kind of surplus we can have to provide income and other support to people for whom transition will be hard.
The economy can grow much, much faster, and people can be upskilled much faster. We should think about the transition period. There’s very much Utopia at the other end, but, but there is a transition period, and people will get disrupted.

Are you talking about guaranteed work or guaranteed income?
India has schemes for guaranteed work. I think we will have to get guaranteed income.
Should it be a universal basic income?

I think it will be absolutely required. I first wrote about it in 2016 that AI will cause great GDP growth, great productivity growth. I ended that piece by saying, we will need to think about universal basic income. It’s very clear to me, it’s where it leads, but there will be enough abundance.
If we are thinking of the world in 2050 do we want farm labourers to be earning farm wages working in 45 degrees centigrade heat in the fields? No, that’s not a job. In my view it’s servitude, but it’s a necessity today, and if we can change the nature of those jobs then it’s a good thing.
The notion of GDP creation and GDP acceleration is separate from the notion of distribution in the west. It’s a controversial idea… everybody wants low taxes. India is in a much better position to tackle it, things like guaranteed work schemes were a bad idea from an economist point of view when they were enacted. They may not be a bad idea in the world of AI, because they are really ways for redistributing income.
How should AI be regulated? There are differing views on this.
I believe there’s some right level of regulation. I was very involved in what President Biden did with what’s called the ‘executive order on AI’, and that was the right level to focus it on.
The technology is in very early stages. Remember, it was only December of 2022 that ChatGPT was released. It’s so new, most people haven’t had time to adjust. And my bet is that in 6-12, months from now, the capability will be substantially better than it is today.
Technology is making rapid progress, but, you know, human beings adjust at their pace, and jobs will take a decade or two to adjust.
If you look at my robot prediction (over a billion robots that will do most of the low-value labour work done today), it’s like creating a business larger than the auto industry globally. How long does that take to put in place? A long time. So the physical world moves at a different pace, even with technology like robots. We just have to recognize that these things have to be managed and managed closely. Regulation may need to protect certain people.
What do you think about the future of entrepreneurship as an individual passion, and startup as a business in India?

The venture business in the US is very different what’s what’s called the private equity business, which is the investing business, where you run spreadsheets and you calculate rates of return. We just bet on big ideas and we bet on technological risk. India’s startup system has been closer to private equity than real risk taking.
If you need a physics invention or an AI algorithm invented, you don’t know you can do it when you start. So, risk taking is very different from the financial venture ecosystem in India. It is less risk taking and more focused on growth investments. I hope we start to get to more fundamental risk taking in India.
Should the government play a role in setting up a large fund itself?
I don’t think governments should be in this business. It’s too dynamic a business for governments.
You were one the earliest backers of Sam Altman. What prompted you to invest in Open AI?

Look, there are so many good ideas. People don’t try them, and when they try them, you need certain types of special individuals to make them happen. These things don’t happen because some academic made a forecast. They happen because one person decides to change things.
When Elon Musk said, I will build an electric car, every major automaker in the world had given up on it. In fact, the department of energy here did a forecast for the number of electric cars in 2035 which Elon Musk (Tesla) exceeded in 2016 itself.
So, big companies and institutions don’t do these things. We have to find somebody who has a vision and can make that happen. Mukesh Ambani did that with Jio. He said, I will lay fibre… I don’t care about the market. Everybody thought it was a stupid 10 billion investment. He just made it happen. So it takes individuals, and Sam was a big part of my belief in AI.
Dependence on technology is also creating a lazy workforce at a time when India needs to reap benefits from the demographic dividend…

The one problem technology will not solve is human motivation. If there’s an easier way, humans tend to take it. I am not sure we should say technology will solve every problem, because human motivations will not change. They haven’t changed for the last 500 years. In the new context, they’re presented differently, but personal self interest has always been the case. Less work and more goods has always been something people strive for.
You have worked closely with Nobel laureate Muhammad Yunus who is heading the transition government in Bangladesh. What are your hopes for him and the country?
I have been to Bangladesh and over the last three to six months, I have been in touch with him (Yunus) and all of his people and writing in support of him when he was being put in jail. I was very much advocating he not be sent to jail, which the then prime minister had wanted to do. I was very much part of the process of trying to give it enough visibility so he doesn’t get sent to jail.
Most of it is through public relations influence. There was a full paid ad in the New York Times. Richard Branson and I and a few others funded that. It was much less with the government itself, much more general public pressure and visibility, which obviously autocrats don’t love.
What is your assessment of his ability to sort out the issues in Bangladesh?
It’s not my area of expertise. I have been there, I have talked to him a lot. I know he is a caring man, and we have gotten to know him over the years and his foundation. But as to the status of the country, it’s very hard to say. As far as I can tell Yunus has said that he has no interest in being in politics, so it is very unlikely he will play a role other than the interim role. He is mostly interested in helping the transition and getting out of the way, and hopefully we will see democratic elections.
On July 22 you had an exchange with Elon Musk on X which was one of the most mature exchanges of views that Musk has had on US politics…
My interaction with him on Twitter was the most rational discourse I have seen. It was not emotional. It was rational. It got close to 14 or 15 million views because it was rational.
Elon is an emotional person, so his comments on the UK riots, I don’t know why he’s wading into that. Obviously, he is not doing it because he has an agenda. He has nothing to gain from that.
On the question of liberal views versus sort of super conservative views. I don’t think he is particularly religious, so it’s not like he’s a religious conservative. I do think there’s a personal side of him that rails against diversity, equity and inclusion, because of the situation with one of his children. I think that’s where it goes into the emotional domain as opposed to the logical domain of what’s good for progress, what’s good for government. He is a very smart guy, obviously.