In the world of AI, Andrew Ng is a name that needs no introduction. Andrew, famous for co-founding of one of the world's top learning platforms in Coursera, along with side other ventures such as Google brain Deeplearing AI and Landing AI, is an adjunct professor at Stanford University and has led AI initiatives at Google and Baidu. Today Andrew is considered to be one of the most influential people in the space of technology. 

Earlier this month, Andrew Ng spoke on NASSCOM's XperienceAI virtual summit on the role of AI in the economic growth of a developing country. While giving his keynote in XperienceAI virtual summit, Andrew Ng shared his view on AI's potential to transform human civilisation as well as the challenges in accomplishing that potential. 

"For the last few years I have been sharing, that AI is the new electricity," Andrew said. "Much like the rise of electricity around 100 years ago, AI is poised to the same," he added. 

In recent years, the world did witness the glimpses of this immense transformative power AI brings, especially in consumer software and internet industries. Today, online processes like web search, and advertising, etc. are transformed by the rise of AI, according to Andrew. He believes the best is yet to come as the potential for AI adoption from industries outside software and internet is more significant, and the next phase of AI revolution starts from the broad deployment of AI in these other industries. 

However, when it comes to AI adoption, especially in the industries outside software and IT, the challenges are numerous.

"The playbook used by tech companies like Google and Baidu doesn't work when it comes to companions from other industries," said Andrew. 

Bridging the gap between proof-of-concept to production deployment of AI system

Even though there are many proof-of-concept projects around the globe in numerous industries, the inabilities of the projects to scarce to production deployment is another major challenge, according to Andrew.

"A lot of conversation on AI today is focused around the Machine Learning systems. If you look at production systems the whole Machine Learning code is just a small piece of the puzzle, there is a lot of other things to be done in order to scale proof of concept systems," Andrew added. 

According to Andrew, there are three main challenges when it comes to bridging the POC to production deployment. These are:

1. Small Data: Many of the industries does not possess Big Data like the Big Tech for training AI system. This can be addressed through synthetic data generation, transfer learning, few shots learning, etc.
2. Generalisability or robustness of AI systems: Today, there are many AI models that work perfectly in published papers, but often fails to produce the same result in a production setting. One reason for this is the quality of data it gets during the research part and the challenge to deal with not so perfect data sets in production settings or large scale deployment. By making models ability to generalise to a dataset other than on what it was trained on is crucial for large scale AI deployment. 
3. Change Management: We humans are change-averse and managing humans to accept and adopt AI technologies itself in their field of work itself is a big challenge. 

The key to large scale AI deployment lies in solving these three challenges as Andrew pointed out. 

India's AI potential

When it comes to the potential of India in AI, Andrew was enthusiastic. "AI has a huge role to play in India, and India has a huge role to play in the rise of AI," he said.

"AI can help Indian economy grow by an additional 15% by 2035", pointed out Andrew citing a recent Accenture report. "India specifically has a lot of potentials, more than what many of you may realise,"

In fact, Andrew's faith in India's potential is based on numerous trends especially with regard to the AI talent India is capable of producing as well as the growing community support in the country, 

"I am excited about the rise of the INDIAai portal bringing together communities and communities really matters," said Andrew about the National AI Portal. He added that trend such as community building, rapid digital adoption, and capacity to produce AI talents in vase scale makes India a perfect candidate for a future AI powerhouse.