India can't afford to slow down in global race for AI: Google Brain founder

India is facing a dilemma: The country must invest in artificial intelligence (AI) to avoid being left behind, but it fears that jobs will evaporate as automation takes over. Andrew Ng, adjunct professor, Stanford University; founder, Google Brain; and co-founder of Coursera, tells Alnoor Peermohamed that India must wrest the initiative rather than rely on multinationals, and it needs to move fast as the “window of opportunity may not remain open forever”. Edited excerpts:

India is in a dilemma where we have too many people and too few jobs and now AI is putting even those few jobs at risk? What strategy should India follow?

I think trying to slow down AI would be a huge mistake. The world is racing ahead to build AI and it would be a missed opportunity for India not to embrace AI. The ethical thing to do is to embrace technology, but also to invest in helping the people whose jobs have been disrupted. Here, I think conditional basic income is very promising, where you give people a financial safety net, by paying them to study. This increases the odds of someone gaining the skills they need to veer back into the workforce and contribute back as taxpayers. Adoption of AI is clearly creating value.

A lot of people are worried about how AI will affect jobs. At first mundane tasks will get automated, then more complicated ones. But can AI be a job creator too?

The field of AI needs a lot more people, that is true, but I don’t think the solution is for everyone to go become an AI engineer. I think the call centre jobs will go away, a lot of driver jobs will go away, and even radiologist will be affected, but the solution is not for everyone to learn to do AI work. There are other job categories—we can’t find enough health care workers or teachers. The job story is actually really complicated. In manufacturing we do a lot of work in visual inspection, and now we can automate the work of visual inspectors, so this will have some impact on jobs. But as it turns out, when you find a defect in a part (we can do that efficiently using AI) it creates a job for a human to go and fix that part. If AI is a substitute, it might reduce jobs, but if AI complements, it can actually create jobs.

Do you think Indian entrepreneurs should look at using AI to solve problems that exist at home?

I think Indian entrepreneurs should work to make India better and even build products for India and ship them overseas. But I think India should try to build a lot of things itself rather than rely on multinationals. But it doesn’t mean India should do everything. Fifteen years ago was a great time to build a search engine, but today it’s not. Today, there’s a lot of AI opportunities worth doing, from agriculture to health care to whatever, and this is a great time to work on them. The window of opportunity may not remain open forever.

What is the method India should use to build this AI prowess? Should it shut out foreign companies or should they invest in education and research?

I would say invest in the next wave. Use the best ideas that exist globally for the last wave but focus on investing in the next wave. I think AI is a category that is very immature, no country has a mature AI community and no country has a mature AI technology. It feels like the US and China have done a lot, but it’s still so early. I don’t think India should write its own mobile operating system, there could be some reasons to do it, but why not take those resources and invest in the future. Take those limited resources and build an AI thing which is much easier, will grow much faster and create much more value for the world and for India.

Whose responsibility is it to invest in and build AI, the government’s or the private sector’s? What can India learn from the US and China here?

I think the governments have a huge role to play, both the central and state governments. There’s a lot a they can do, from community building to investment in research and education to public-private partnerships as AI moves forward. Take for example self-driving cars, that would need regulatory changes, so the government working with private parties will make self-driving cars happen faster. Companies have to invest in workforce development. You can’t just wait for universities to graduate more people (although universities should graduate more people), it’s not enough. They also have the responsibility to help people whose jobs have been displaced. For individuals, lifelong learning is the new constant. I think that individuals who constantly keep learning will create a better future for themselves and their families.