In next 5 to 6 years, AI to make significant impact in healthcare and IoT space: Beena Ammanath, HPE

A few years ago, the industry was in the narrow intelligence stage and it was expected to take a few more years to get to super intelligence. What is the current status?
I think it is still narrow, however we see certain industries combining narrow intelligence to do more complex tasks. For instance, in a self-driving car, there are different components of AI that come together that make it a self-driving car. Also, when we think of general or super intelligence, we imagine a machine doing everything that a human can do. However, for this to happen, we require emotional intelligence, which I don’t think we have the underlying models or programing languages for it yet.

There have been talks about the development of AI and most IT organisations are pushing it in the market. Still, why is there so much of apprehension in the market to adopt it?
Today, we are using AI to automate existing processes and defined tasks, which are essentially done by humans. We haven’t fundamentally thought about how do humans and machines work together well. We need to understand which tasks should be handled by humans and which should be done by machines. For instance, we hear a lot about using AI to detect cancer at the early stages, but the actual question should be how can we use AI to prevent cancer?

How should challenges related to the adoption of AI such as data complexity, ethics be addressed?
Discussions on ethics are taking place at a global scale and we are at the forefront with that because we have visibility. Also, ethics are geo specific because what might be right in one country could be wrong in another one. As we bring this global perspective, we try and define the framework. However, the number one problem today that we see is of adoption. For instance, you can build the most accurate AI model, which gives a 99.99 per cent accuracy. If such a model tells you that a machine in the factory is going to fail in the next 24 hours but if a factory floor worker refuses to trust this model, it is a failed project. So it is important to put design thinking upfront, and ensure that the users of this technology are willing to adopt it, and accept the feedback.

There are also times when people are fearful of AI taking away their jobs, however this is not true. In fact, what we see in some of the more advanced areas, new jobs are getting evolved. For instance, there is a job title called ‘sensor cleaner’ for cars whose job is to periodically clean all the sensors on a car. We also expect to see growth of jobs like AI architects, testers and people who are giving feedback. As the skills required are slightly different, we will see a lot of initiatives around reskilling and upskilling coming up and at HPE we are helping our customers with it too.

Do you place a significant impetus on creating a right partner ecosystem so they can tailor the right AI solution for your customers?
Yes; for instance, in our predictive maintenance offering, we didn’t even build the machine learning piece. We have a recommended architecture at the bottom which has connectors. So if you have a factory and your data is on a cloud, we have connectors to it from our partner called Dataiku, a start-up whose models are very much aligned to factory machines. Then on the UX side, out of the box, we offer Spotfire, but we also have connectors to Tableau. My team continuously evaluates start-ups and delivery partners, so that we don’t have to build everything from scratch.

Where does India stand on in terms of AI adoption?
I definitely see India investing much more in AI and the adoption is at a much higher rate. Companies in India are willing to experiment more and are pretty aggressive about it. We’ve seen a lot of traction in the IoT space with smart cities and smart parking. However, we struggle when we need to justify about automating something which can be achieved by cheap labour. Also, while India is good in applied AI, it lacks in AI research. Unless the country invests heavily into its universities and academia on the research side, it will always be dependent on the research that’s happening elsewhere. However, I’m pretty confident that there are really good companies and start-ups that are doing innovative work in India that can go on at a global scale.

Given your experience in the industry, how long do you think it will take AI to be the order of the day?
In a mindful way, it will be in another 20 years at least. However, in every aspect, it depends on the industry. In the next five to six years, AI will make a significant impact in the healthcare and IoT space and this will drive the next wave of mindful AI.