The age of machines is certainly upon us, but not as a Sci- Fi. AI should be interpreted as AUGMENTING INTELLIGENCE, helping the human do better at what he or she is doing, and with balances and checks in place for security of near future generations, AI could make a great team partner, says Ullas Nambiar, VP Innovation & Head of Zenlabs
AI is not an emerging technology anymore. Where does it stand today?
AI was never an emerging technology, in fact a whole bunch of technology has already been a part of AI – and it has always been consumed as machine learning etc. AI is essentially everything. The perception of AI as we see in sci- fi movies, are 30-40 years away. What AI is like an IBM Watson or a Deep Blue, essentially a smart system that is able to understand a certain space. It’s basically a learning system. AI has a long history, as everyone knows- starting from the 60s, chess players have been playing with these. The growth has been slow, and it has been building up for a long time- Eliza for instance, the Mars Rovers managers- managing system are also AI. Robots have even been working on shop floors. This is evolution of AI in terms of machines being managed in a very controlled manner. It was, thus, never an emerging technology, maybe an emerging buzzword- in 2017, startups are offering AI, and VCs are talking AI.
Today, it has evolved for larger good; there has been some understanding of what AI could offer for the human good, last year. From the business perceptive AI became mainstream only around last year, though academia has been doing it for many years. The popular imagination has been caught by AI now.
In your opinion, what industries will AI affect the most in the short and mid- term?
In the manufacturing industry a lot of the automated processes on the shop floor are being done by robots today and Amazon’s warehousing has a lot of robots which move crates. So I would say repetitive jobs with a basic skill set are being robotised. Those industries will definitely get affected. The basic bottom of the pyramid activities which can be automated will be taken over very quickly. While managers cannot be replaced because their work is not a process, lawyers can get potentially eradicated, because if you have rules with very strict parameters, robots can do those jobs.
In short, there is no industry that won’t be affected. Drones, for example, can be used for mining, agriculture, based on our imagination and ability to use for beneficial activities. Of course it will lead to some job losses, but new jobs will develop, depending on what we have seen in the past.
Is security an issue with AI apps? They use data and so, what kind of security measures does Zensar recommend/install to fight risks?
In any tech offering, security is an issue. We don’t give enough credence till an event happens. It has not become a focus area because, what is public and what is private differs accord to the demographics involved. Having said that, it is an area of critical importance. For instance, in healthcare, a lot AI powered tools will put a lot of patient data into systems. So that notion of privacy has to be very clear, in terms of who has accessed, and once it happens, it can be tracked back. Traceability has to be built into the systems.
Another aspect of security is a malicious one. So if somebody can induce a malicious attack, say on a human being, we need to have those checks and balances in place, as we go towards more and more AI powered systems. As of now, we have Asimov’s Laws that will protect us, but soon we will need to go beyond these three rules.
With automated systems in place, there will be no boundary of business and personal data. It’s not just human information, in terms of AI, it’s the fact that you could control a large machine and make it attack a human being. It can control the missile and send it out to some other country. Those are new and bigger threats compared to privacy of data, which is already an issue, faced by current traditional systems.
We will then have guarantees in place so a system you have built cannot be controlled by a malicious intent, by a third party. Those have to be built in when you build the systems.
How do you think technology like AI can help meet larger goals like Digitization of India?
The frank answer is, for AI to be in place, digitization will bring a basic infrastructure, bandwidth, networks etc. So it will brings the roadways, and AI will be the trucks that will run on that. You need digitization first. To have pipelines, you need that connectivity to be there, to place smart systems, aiding each other. If systems are not smart or connected, with the ability to be on a digital bandwagon, then AI is useless, it’s just a lab toy. It’s a lockstep. We need to first lay out our 5 G networks, broadband etc, then on top of that we can put all those tools that the smarts AI can bring.
In five years, what industries do you see becoming completely dependent on AI?
Enterprise adoption is questionable, it depends on each part. Like I said, repetitive tasks will be taken up, AI is better than humans for sheer scale it can manage. In 5 years, every industry will get effected, some more some less. Even IT, because certain kinds of coding can be automated. Projects are already on the way.
However, the union of human and machine that will happen will not be replacement. AI should be interpreted as AUGMENTING INTELLIGENCE, helping the human do better at what he or she is doing – a team of a human and a machine working together. In that case it’s a great partnership. So in five years this should happen, it’s a learning curve but we will reach there, because if competition is doing it, you will also be forced to do, it’s a market place. Competition will make sure that everybody starts adopting some more, some less, but in 5 years, everybody will have some basic elements of AI. It will be smart automation, where somehow decision making can be moved from human beings to machines.