1) Can AI be a threat to humanity in the near or distant future?

I'll start by saying that in our school days, we all used to have an essay on science whether “Science is a blessing or curse”. So, we'll see both sides of it. However, in my honest opinion, I truly believe that it's not at all a threat to humans. It will definitely change the way things are getting done. Currently, I believe as one of Andrew's quotes says “AI is the new electricity”. Like how electricity has revolutionized 100 years ago, AI will also transform every industry in the coming decade. However, it's not set at all to humans as a threat.

2) In today’s time, terms such as big data, machine learning, and AI are often used interchangeably around AI systems. Could you simply these terminologies for our audience and explain how they are connected?

I guess these terms are often used and people do get confused. However, I'll just start with the very simplest form of how we can perceive AI, as it can be seen as a specialized field. A field that combines computer science and datasets for any kind of problem-solving. It can also be seen as a simulation of human intelligence processed by computer science.

Now machine learning is actually a subset of AI. You can see it as a subset of AI wherein the model gets trained on historical data which actually helps machines to make better predictions quite similar to humans.

In machine learning, you can see this area divided into two broader categories. One is broadly supervised learning and the other one is unsupervised learning. In supervised, we generally use historical data to predict any predefined outcome.

Basically, a lot of independent factors whereas in unsupervised learning hidden patterns are uncovered to understand the segments that exist in that dataset. However, the backbone of this AI and ML is actually like big data.

Just to understand this in a very generic term around a couple of years back people start defining Big Data as any sort of data that can be classified broadly in 3V’s characteristics. These are high volume, velocity, and variety.

If any data set is having all these three characteristics, we generally call it big data. In the digital era, we have a humongous volume of data and so big data is largely the data that is huge in volume, velocity, and variety in the stream at which the data is coming in on a real-time basis.

We connected across various digital assets through which you get data in terabytes and petabytes. So, big data is all about having all these characteristics. It also contains both structures and as well as unstructured data sets so that's how you can see these three terminology.

3) How do you explain these technologies to business stakeholders who are not from a tech background?

Most business users are actually not keen on knowing the detailed technicalities around data insights. Like a long time ago, I built one Advanced statistical model around the fraud monitoring model for one of the health insurance companies. When I was in the boardroom and I was trying to explain the complex equations to the business stakeholder but I found it difficult because the end consumer was not tech-savvy.

They were finding it very difficult to follow so to overcome their problem, I translated the findings as a layman for them. For instance, showcasing the simple trends of highlighting the reasoning behind the same like “why the trend is going up and down” and then once people start relating to it they get interested. They ask the right set of questions so once they generate complex data inside I would say we must simplify it for the stakeholders so that they can accept or appreciate or make the right set of inputs to make it more actionable. It is because, at the end of the day, analytics outcomes should be able to make an impact and improve the set KPIs for the organization that I guess you should focus on.

4) To further elaborate, what is the impact of data on business, or how exactly data is the new oil?

Well, every industry is sort of collecting tons of data directly or indirectly. Now the companies have realized and started seeing data as an asset for them. At some point in time, there were a lot of companies that were storing data but they were not utilizing it. They were not even sure how to use the data. Now, the awareness of AI analytics has increased. This is why companies have started investing in a sort of analytics “Center of Excellence” as a key vertical for an organization.

In today's world, every organization like I said is sitting on tons of data but the most important is that companies are starting to see this data as an asset to utilize it for various purposes. This is where big data analytics will be, and I personally see it as one of the key cornerstones of making analytics a success in a company.

If the company is not having a focus on big data analytics then they might lose the battle to make an entity a success. Every company keeps on working on various use cases which can be solved through a mix of both big data. Also, big data and machine learning and that's where these models and everything play a significant role in automating a lot of their intelligent and repeatable tasks which can otherwise lower the overall application of an organization. So I would say that every organization has a bunch of use cases. It's all about awareness and setting up a team and some focus around that and eventually, the company will start seeing the value of it.

5) Adding to this, given the growing demands for data analytics, what are the key skills aspiring data analysts should have to work in the AI sector?

I will say for any data analytics expert, the technical skills these days which are of very high demand is SQL (structured query language) which is anyway the key skill. I would say followed by data management skills, data visualization skills, and of course to be a data analytics expert you should be statistically savvy, econometrics, etc.

I guess these are the key technical core skills. However, one of the two most underrated skills, I would say is critical thinking and business acumen which people generally foresee. As a young generation who are aspiring into data analytics, they focus on SQL, they focus on data management but they don't focus on critical thinking and business acumen that's where the key difference is. If you really want to become an analytics expert and be successful, this is the key.

6) How do you think AI and machine learning systems are likely to change business operations in the next five years?

I would like to highlight five key areas where the analytics field will flourish. It will include of course include artificial intelligence, data democratization, augmented analytics, data as a service, and data governance.

These are the key areas where analytics would be evolving in the coming five years for businesses. To stay ahead, I would strongly recommend that they must heavily invest in analytics as a center of excellence whose primary focus should be to solve various data silos which exist in a sort of company. And try to create one data platform which has all kinds of data in one place and respective business KPIs are enabled on top of it. This at the end of the day will eventually lead to this data democratization. Once the data platform is achieved company should start focusing on augmented analytics wherein their machine learning models can process and generate insights automatically.

This data can be directly consumed by the various business stakeholders once data is there then and there. No data ambiguity in there like everyone is following the one set of data now and serving the business stakeholders on top of it. This will help business users to understand their context as relevant questions are uncovered and insights are gathered more quickly as well aligned.

7) What are your thoughts about creating an analytics excellence center, what does it entail about the capabilities, the workforce, and the roles?

It's a team of people having different skill sets so this whole analytic center of excellence has people from a data analyst background, business analyst background, and even storyteller. It’s also a big skill that these people are hiring especially storytellers.

In our analytics center of excellence whatever effort is being put in creating the model, if you are not able to translate then success can't come. Therefore storytelling is then statistical modelers,  and data scientists so there are different titles. In fact, I also have worn all these caps in my like two decades when I started my career as a data analyst then moved to business analyst wherein I started using Advanced dashboarding tools then moved to statistical modeling wherein started with logistic regression to neural networks and then move to this big query platform.

We are utilizing all the AI models and then deploying those on the production so one analytic center of excellence has six-seven different kinds of spaces. What I am largely saying is that people from different backgrounds are joining. It's not just a team of computer Engineers like I said, there are also designers. It is because at the end of the day like analytics at some point in time will give some signals. Now, it is the creative team that helps make it more Effective.

Adding to it, analytics help them to augment productivity. This is how we can see this as a central excellence in some companies. It could be just a starting point then it could be this business analyst and data analyst. In fact, there is a data engineering team who is establishing the whole layers of data.

8) Talking about data democratization, what is the underlying cost that's being paid by me or any other user for using these free AI tools and services?

I would say that whenever a user is interacting with a digital platform, it is giving a lot of its own data. It is giving a lot of signals about what it is, what he is doing just to give you the sense around the Times Internet. So we have a news platform wherein everyday user comes in and read different kinds of news articles across genres and so we have around 600 million odd monthly active users who come on our platform but every user is having different behaviors, different patterns, different tastes, bias, etc.

Of course, they are there because the platform is free but they are giving us a lot of information about what a different set of users are interested in seeing and this is helpful for us to sort of segment the data. On that basis, our major revenues are coming from the advertiser so while selling our offerings to the advertisers, advertisers came to know that you are saying that you have 600 million odd views but for my product who are the top prospects to serve? This way we collect huge behavior data around the users which help them to pitch the right set of people.

9) Can you walk me through what is involved in the process of obtaining consent from users when it comes to our data?

Data privacy and ethics are basically crucial aspects of building and maintaining customer relationships, especially in the digital age. To obtain consent from the user, there are a few things that every organization keeps in mind and that's how we also follow.

First of all, understand the legal framework depending on the geographies in which you are operating. For example from which different geography your customer belongs to. Any respective regulations that are available in that particular market, we should have adhered to particular regulations for instance GDPR, CCP, etc.

Those kinds of regulations are there so we have users coming from European countries visiting our platform. We need to be compliant based on that understanding. The legal framework is one of the key aspects. Another thing to learn is communicating very clearly and honestly like “what to collect?”, “why to collect?”, and “How to use it?”. Also, whom you share it with let's be very transparent to that so that users trust.

Trust is the key right so once the user is very clear on their understanding of what is being collected for and for what purpose then they respect that. Of course, respect customer preferences and expectations if someone is not interested in sharing so that kind of respect should be there in those processes in place.

Last but not least to accomplish that security and accountability need to be enabled at a very green level. Again, the customer trust depends upon that so whatever consent the customer is giving follow that diligently and make sure that things get to the right side of expectation.

10) In your expert opinion, how do you strike a balance between experimentation and ethical responsibility for a business when it comes to AI and analytics?

I will start by saying that see trust is one of the most important factors in any sort of relationship. Analytics can flourish in any organization if mutual trust exists between all the stakeholders for companies to achieve.

I guess ethical and transparent AI is necessary to make the AI trustworthy. How can we achieve this probably, well, it can be started by increasing AI awareness, clear communication, and AI algorithm.

We should try to demystify in explainable terms as laymen. For instance, how AI is making decisions for people. Any doubt in any stakeholder, we should clearly highlight how AI makes all those decisions, and last but not least, I would say they follow the rules and regulations that are applicable for any particular organization because that's paramount.

11) Could you walk me through how exactly an AI system let's say ChatGPT approaches the process of decision-making?

At the end of the day in the digital era, we have huge data. Data is coming up and we are now converting the data into signals or you can say that in information there are tons of variables that can be created through those data sets.

Once the things are there then there are statistical techniques available to correlate things and that's where for example in our brain they know how our brain works. There is a neural network that exists in our brains.

The neural network is the same kind of terminology being used in data analytics which uses the same kind of processing around neurons, data signals, and bases. The outcomes come in and you can clearly see that if let's say you can also see the accuracy and the prediction of the model by stating. For example what the AI model is saying versus what human intelligence is there and then that's where the accuracy and prediction comes in.

You can see that it totally depends upon, “How much big data you are infusing?”,” How big signals you are infusing?”, and “How the best model you are using?”. So like ChatGPT and other AI technologies are based on that data. They view data and it gets fed into these systems, and now considering that the tech cost & processing cost is very cheaper, the models get trained on the huge data sent. Due to this, the ChatGPT is capable of answering anything intelligently and whatever you are asking.

12) AI replacing human jobs is one of the conversations. In your opinion, which sectors are most likely to take a hit because of AI?

I would say jobs that include redundant and repetitive tasks will be replaced as AI will be able to do it more smartly and efficiently. For example just to give you a sense like customer service can be majorly improved with the use of AI and it applies across industries.

We will see a lot of involvement across various sectors for example in a media company we can use it for advertising content creation and technical writing. All that kinds of use cases are there and applicable to media companies, for the legal industry, you can see that there are legal assistants that are coming up who are actually helping the lawyers. In fact, there are things for the customers to fast-track their cases, even prioritizing what to focus on and what to not focus on. However, at the end of the day, I would say that AI won’t be fully able to automate these jobs because it requires a fair degree of human touch. For instance, human judgment to understand what a client wants will definitely improve the productivity level of the people.

13) As an expert, I was curious about your perspective on how AI will be reshaping society in comparison to what is projected in web series and movies. How far-fetched are they from the reality?

I'm not sure how it will change but I would say in the next decade, it will be utilized as a strengthening tool or sort of strength for different kinds of verticals. As I said AI is the new electricity so there were people living before the electricity era but now the electricity has changed everything.

In the next five to ten years, the kind of workforce, and the kind of verticals across the companies might reshape across industries. There might be different sets of rules, and profiles because AI will take the jobs which were related to the repetitive tasks or something like that.

The people will be more prone to do upskilling and reskilling for different titles. For example, when the calculator got invented at one point in time people started saying and I read this somewhere that what would be the role of mathematicians. However, people started realizing that it augments things and does not just replace anyone. Therefore, the idea is it's difficult to change the whole ecosystem and that’s how I see it.

14) Do you have any particular tools that you would recommend to advertisers in terms of enhancing their creative capabilities?

These days like we have OpenAI (makers of ChatGPT) is everything and open source technologies. We should have the right sort of infra in place that's what I would say. Whether you use AWS or GCP, you need to have the right set of infra and the right set of capabilities.

There are multiple courses for any language, open-source libraries are there, and one can just create a complete data science lab. We also have established a data science lab at Times Internet wherein the data scientist can just use the tool.

However, I would say Jupyter Notebook is what people generally use to create any kind of model of things, and these days, we are also utilizing Bigquery i.e. a Google product. Basically, we use GCP and Bigquery a lot because one is used for data processing and the other for data aggregation. On top of it, Bigquery is good and this is a Google product that offers a lot of advancements.

Google has unleashed its whole AI model capabilities to use. It of course is not for paid business but it generates huge value for us. Therefore,  I would say every company can have a different path. I won't say that you take only this path depending on the company size and companies data volumes.

The kind of focus that a company wants to put in like I said analytics center of excellence must be there in any organization. It could be at a starting level for some organizations to the maturity level but if the CEOs and CXOs are not focusing on that then they are two years or three years down the line in comparison to their competitors and their competitors might have a better advantage on a lot of use cases.

15) Lastly, when it comes to AI and analytics, what trends you are excited about and looking forward to?

As an analytics leader in the company, I focused on solving various business use cases which are relevant to the company. Use cases that can help companies generate more revenue. Based on that a lot of advances are anyway happening and so translating those use cases into AI-driven projects is what I am more focused on.

For example, I'd just give you a simple example, in the advertising world for the advertisers the key KPI that they focus on is the CTR. It is the click-through rate. So any advertisement comes in what the CTR you are giving me. So, of course, one is depending upon a lot of factors related to the advertiser, brand, and position they have.

To get written, those are factors that we cater to in our pricing model and everything on top of it. Whether it is creative in itself, what are the signals, what is the color of the creative rather but the content is being written for the creator. So all these advancements will augment the capabilities of CTR. For instance which creative works or we create a new creative for an advertiser.

For example, an advertiser has given us a brief and given us just one sample creative. So instead of running one creative, we run 10 different creatives and then AB tests those. This way, I can keep talking about different use cases but this is one of the use cases that helped us to sort of use the AI and advanced techniques, the image processing capabilities, and all that we are able to achieve with that kind of thing.