Over the last decade, the pace of technological innovation has increased dramatically. In addition, breakthroughs have grown exponentially in the domain of information technology, particularly in artificial intelligence (AI) and machine learning (ML). These technologies are working to boost human efforts, to take on just about everything from the quality of life to cybersecurity, and so on.

Aishwarya Srinivasan, AI and ML Innovation leader at IBM, talks about the future of Artificial Intelligence and shares her journey in the field of Data Science.

Continue reading to find out more!

Q. How much has the field of Artificial Intelligence and Data Science evolved in the last 5 years?

In the last 5 years, we have seen major advancements in the field of data science & AI. The field has definitely been growing exponentially.

Some of the major breakthroughs that we saw in the past 5 years were with DeepMind’s reinforcement learning model Alpha Go learning an entire game from scratch using machine training and beating the world champion, we had a breakthrough with new hardware – TPU which can process deep learning models incredibly fast, we had the invent of transformer models which brought in a whole new level of how NLP models were built, and of course deepfakes which could create things which had only been a fantasy.

With all the advancements, another major thing has been evolving- ethical aspects about building these AI applications. A lot of it started with the Cambridge Analytica scandal when people started questioning what the major tech companies are doing with the data being captured from the users, most importantly, do we even know of what is being tracked.

Governmental agencies, privacy experts, and ethics researchers got together to analyse what is happening and started creating awareness about what is going wrong and what could go wrong if things are not done in a more governed manner. This is when we started seeing a lot of initiatives around responsible AI. 

Q. Which industrial sectors have been the primary areas of focus (for AI) in the last decade?

Based on what I have experienced, the technology industry has been on top of all inventions.

They have been building the right kind of solutions for other sectors while helping them automate processes, build data-driven decision pipelines and help grow their business using advanced machine learning and data science technologies. 

Q. Which new industrial sectors are expected to be the major benefactors of AI this decade, especially after Covid-19?

There isn’t a single industry that has been transformed with AI technologies. Post covid, digitization has revamped every industry across the globe.

Businesses and people have found a new way of functioning and living lives. If we were to just look at how technology has been adopted in the past two years, it is incredible. 

Q. What are some of the major challenges the evolution of AI is facing worldwide?

With AI becoming a global technology and coming into the hands of common people, a major challenge is to have a governed and trusted means of building these applications. 

Q. What drew you to machine learning? Do share your machine learning journey so far.

I started my education in computer science and naturally got interested in data science. To have better industrial experience, I did several internships in wider areas like time-series analytics for macro & microeconomics data, building chatbots, building demand forecasting models, customer satisfaction metric estimation, social media sentiment analysis, handwriting mapping using OCR etc.

With each and every project I was exploring my domain interest in the space of data science and at the same time was sure that this is the field I want to pursue my career in. 

Q. Do you believe that AI is still not being used to its maximum potential? If so, what can be the reason for the same?

Yes, I do believe that AI is not being used to its full potential. On one hand, we see great advancements from research, but on the other hand, only a very small portion of that work is being used in production.

The reason behind this could be that for organizations to put any model into production, there are many checks they need to adhere to like model explainability, model governance etc.

Q. If any organization is planning to use AI for its products and services, what’s the most important thing they should keep in mind.

The most important thing to keep in mind is the economics of productionalizing the technology. The organizations need to understand the cost of building the models, scaling the model, replacing the other process with the new model, understanding what business value it generates compared to what’s existing, and having an estimate of how it would run in a month, 3 months, 6 months and so on. 

Q. Many people believe that technology (AI in particular) has been the major reason behind many employees losing their jobs. What do you feel about this notion?

I don’t believe that AI is taking away any jobs, it is a shift in skills, like any industrial revolution. 

Q. For anyone who’s aspiring to make a career in the field of artificial intelligence and data science, how should he/she start! Can you provide some helpful resources?

Sure, I have prepared a roadmap for anyone to use to get started in the space of data science. https://youtu.be/f5Yd5zo1IdM

Q. What are some of the most important (non-technical) skills that would help someone succeed in the fields of AI/ML/data science?

Some of the most important skills that I feel would be most helpful is having a collaborative mindset, and secondly identifying your USP and building your brand.

A person needs to be open to exploring more opportunities even if they find themselves in an uncomfortable situation, comfort and success don’t go hand in hand. 

Q. What are some of the biggest myths people have about AI/ML/data science?

The biggest myth that people have about AI is that it is some crystal ball or a magic box, which it is not. It is built on pure mathematics and statistics. AI or machine learning algorithms aren't going to solve major business problems on its own. For the solution, we need to craft and tailor the algorithms to fit the problem. AI shouldn't be looked at as "pure solution", rather it is a "tool" to solve the problem.