VMware CEO On Generative AI, Jobs And Criticality Of AI In India’s Global Ambitions

According to the latest report by Goldman Sachs Research, Generative Artificial Intelligence (AI) could contribute as much as 7 per cent or about USD 7 trillion to the global GDP over the next 10 years. The technology is also projected to lift productivity growth by 1.5 percentage points over this period.

So massive is the potential of Generative AI that its relevance is now being discussed in quarterly earnings results. On Thursday, Infosys CEO Salil Parekh during the Q4 FY23 earnings press conference mentioned the use of generative AI capabilities multiple times for clients and within the company as well to raise productivity. Even Amazon CEO Andy Jassy, in a letter to shareholders, leaned heavily on making huge investments in Large Language Models (LLMs) and Generative AI. Similar efforts are being undertaken by companies and businesses across the globe and India, irrespective of their size. 

A recent TeamLease Digital report predicts that AI will add USD 450-500 billion to India’s GDP by 2025, accounting for roughly 10 per cent of the country’s target of USD 5 trillion GDP. 

With such massive anticipation around AI technology, BW Businessworld’s Rohit Chintapali spoke with VMware CEO Raghu Raghuram in Mumbai to understand the pace of innovation in AI, possible job losses due to the technology, the need for AI policies and what India’s approach should be with the technology. Excerpts:

Should we hurry up with policies around AI?

There has to be a significant engagement by the policy makers. I don't know if there is enough understanding of the pace at which this AI space is developing. Eventually, we do need policies but it has to be done – given the nature of how fast it is evolving. Moreover, it has to be done in a very thoughtful way by people that have a fairly deep understanding of the issues.

A recent Goldman Sachs report mentioned that as many as 300 million jobs could be lost due to AI systems. Do you think such job losses could happen in next 10 years?

I don't think so. Generative AI technologies require a human in the loop for a lot of activities because it can make spectacular errors. You need human judgment. But even as this develops, it could completely replace human functions in certain areas but there will be other newer things that will be created – that's always been the history. I mean, they predicted that accountants will go out of business when Excel was first introduced. Now, there are more accountants in the world, per capita, than there were at that time. 

So, the nature of jobs will change?

Yes, there will be more job categories or jobs that we do not know anything about right now. For instance, nobody knew that a social media influencer/content creators would be a job category 20 years ago. But the dystopian scenario of everybody is sitting around with nothing to do and the state has to feed them, I think that is a little bit farfetched.

Zoho’s Sridhar Vembu recently said that AI was going to make coding easier and that companies wouldn’t need armies of developers in the near future. Hence, he was not very positive about creating more jobs.

I disagree. 

Could you explain?

Think about it. Why has ChatGPT acquired a 100 million users in two months? It is because you can speak to it in English. So, English has become the programming language of the computers. But what is the biggest backlog for any company that you walk into? It is the number of developers. They all say, I do not have enough developers to write all the software I can think about or my customers need. So, when that friction eliminates, you will still be developer. 

If you can specify a sequence of procedural instructions to a computer and say, “do this, this and this” and “here is how you make it” – you are still a programmer. It is just that you are instructing the computer to do a sequence of activities, except, you are doing it now in English. AI systems are, in turn, just doing the hard work of generating the code, just like compilers did before. 

So, what Vembu said was that productivity of developers could be improved and, so, the software developers will be affected job-wise because you would not require as many people to generate code.

True. However, there is a lot more code to be written at this point. It's not that there is a finite amount of code to be written, right? There's a lot of uninteresting code that every developer has to write and put a program into production. Co-pilots could automatically create all of that. And those kinds of things will happen plenty. But then, developers, in turn, will be able to do more code in a day. 

What are your opinions on companies releasing services to the public with an “experimental” disclaimer? Is it safe to do that, given the potential of Generative AI and how people may find it hard to understand outcomes/false outcomes?

Let me address this with an example. When the first App Stores came along, there were a ridiculous number of bad apps. There still are. But once you tried them, it gave you an idea on the good ones and better apps. So, the nature of new technology is such that there always will be experimentation. It is great because that is how you get advancement.

But powerful experimental services are justified?

As long as you put disclaimers, yes. When technologies are moving so fast, it is hard for everyday people to truly understand what the power of this thing is, or, what it can do or what it's good for. But the only way you can bridge that gap is if you actually use it. Otherwise, you can't form an enlightened opinion. 

You’re saying that the users (informed or uninformed) are genuine stakeholders in the advancement of technology like Generative AI as they could learn experientially and form an informed opinion?

Correct. Imagine, at some point, the government wants to introduce a new policy regarding the use of AI How would you know whether it is a good policy or not if you have not used it before? In a democracy, you get a say. So, the more people use it and more they will be enlightened about it, the better it is for both the advancement of the technology and the advancement of policies regarding the technology. 

Are there any limitations? Should we be wary about anything with AI development?

The companies that are building these new AI models and developing the technology, they understand the power of these things better than most people like you and me – who are on the outside. Hence, the onus is on them to both educate and put the safeguards as needed. That is where the higher sense of responsibility needs to come from. 

Do you think countries like India should have AI-first policies to foster innovation at a national level? Because monopolies could easily form on the global stage, especially, with the US and China being far ahead in AI innovation? 

I certainly think this will become a contentious topic in the future. Because national competitiveness is affected by the degree of utilisation of machine learning and AI models. So, certainly governments like India should be investing in AI – not to build giant models which will compete with the ‘latest and greatest’ from OpenAI or Google – but they need to foster the creation of these models and the usage of these models as part of the national infrastructure. This could be considered critical national infrastructure.

You mean, something like the India Stack?

Yes. India Stack was done with the support of the government. You need similar efforts here because this thing is going to be one of the most significant determinants of technological superiority among nations. And it's going to affect all forms of competitiveness, including economic, military and more. Hence, it's super-critical for each country, certainly a country like India, as it aspires to be the top three global powerhouses in the world. 

Also Read: Our Value Proposition Is Even More Attractive During Downturn: VMware CEO Raghuram