Swami Sivasubramanian,vice president of AWS (Photo/NBD)

Sept. 19 (NBD) -- The 2018 World Artificial Intelligence Conference was held in Shanghai on September 17-19, gathering industrial leaders to discuss topics including man-machine integration, AI computing algorithm, infrastructure building, smart chips, big data and AI investment.

Swami Sivasubramanian, vice president of Amazon's cloud computing platform AWS, had an interview with NBD on the sidelines of the event.

NBD: What do you make of AI application?

Sivasubramanian: AI or deep learning is not new, and deep learning can date back to 30 years ago. But the reason why deep learning has not been widely applied is that it needs huge amount of data. Secondly, it really needs huge amount of computing capacity. Thirdly, it really needs a good machine learning practitioner.

Historically, the ability to run this huge amount of storage system and huge computing data centers are only available for big Internet companies but not available for many startups. Now that's the very case where cloud can bring about changes. You can be a small startup, and you can actually build your AI systems from scratch. It is a you-pay-as-you-go, because it is really cheap and it is the mostsecure data storage system.

NBD: Gartner estimated only 4 percent of enterprises have already developed or invested in AI applications. What is your advice to those companies who take a wait-and-see attitude?

Sivasubramanian: To those companies that haven't adopted AI technologies, I would say that AI is such a transformative technology that helps enterprises upgrade supply chain and improve customers experience. By adopting AI, they are going to improve revenue and decrease costs.

That's why many companies' CIOs, CTOs, and CEOs are already asking themselves and their teams what are their AI strategies. So it's not a case of whether one should use AI, but when and how can someone get to start their AI strategy. And the application scenarios are growing in sectors such as healthcare, fashion, real estate, retail, and auto-driving.

NBD: Can you give more details on future plans of AWS' Shanghai lab?

Sivasubramanian: We have a saying in Amazon, 'every day is day one.' Everyday is just like you are starting a new revolution in the internet era. In terms of the Shanghai lab, there is so much research need to be done, for example, what AI systems need to be built and how they can be explained in terms of what actions AI system is taking.

There is actually something we can explain internally to business stakeholders, and externally to laboratory regulators. How can you deploy machine-learning with database technology? There's so much research to be done on the fundamental level. 

Email: tanyuhan@nbd.com.cn