- Managerial jobs and tasks that are repetitive in nature will be displaced
- The ability to learn new skills will be critical for individuals who want to stay relevant
- Companies will need to devise new ways of training and assessing the skills of employees while countries must develop a learning ecosystem
- Work will be more contractual in nature and deep technical skills, creativity and learnability will be at a premium
Below is an edited version of their conversation
Knowledge@Wharton: Artificial intelligence (AI) and robotics are likely to have a major impact on work and productivity over the next decade. How do you think they will affect how work is done?
Srikanth Karra: Predictions regarding AI and robotics have been there since the ’60s and ’70s. I still remember the day when IBM supercomputer Deep Blue beat world chess champion Garry Kasparov. That was around 1997. That’s when I realised that machine power is here to stay. I would say that with the exponential increase in computing power over the last couple of decades, it has now reached a critical mass of industrialisation
In other words, billions of gigabytes of data on machines, process automation coupled with machine learning and the new things which we hear, like deep learning abilities and algorithms, have made predictability and pattern resolution easier and more accessible. This has undoubtedly increased productivity exponentially. Work will shift from repetitive tasks to more creative tasks.
For example, architecting creative solutions in all domains, entrepreneurship, art, governance and teaching are some of the areas where I see more action. Work will be more contractual in nature and deep technical skills, creativity and learnability will be at a premium.
Knowledge@Wharton: What are the kinds of jobs that will be displaced because of automation and AI, and what new kinds of jobs will be created because of these trends?
Karra: Tasks which are repetitive in nature will get displaced. For example, accountants, hotel receptionists, software testers, analysts and financial reporters updating data and creating reports from data. Medical diagnostics is another area where there’s going to be tremendous displacement. Assembly workers, undoubtedly. Legal and financial advisory, middlemen and brokers. These are some that come to mind right away, and there are many more.
If you see the nature of all these tasks and jobs I’ve talked about, they are all repetitive and are being displaced by digital technologies, including AI and robotics. However, there will be many more jobs which will be created. I call them “adjacencies.” In our history of automation, whether during the Industrial Revolution or the many other automation-innovations, there was never complete disruption of jobs, but there were adjacencies which were created.
There is no doubt in my mind that the nature of disruption which is happening now, particularly the technology-based disruption, is going to create a bigger disruption as compared to the Industrial Revolution or the automation of work and assembly lines. At the same time, there will be adjacencies of jobs which will be created, which I’ll talk about as we go forward.
Knowledge@Wharton: If you look at these kinds of disruptions globally, in which regions of the world do you think the impact will be felt first? And how do you expect the trend to evolve over the next decade? Also, in percentage terms, how much work do you think will be displaced by automation globally?
Karra: In my view, the developing countries, where there are large, unemployed populations and which are just about to embark on industrialisation, will be hit harder. Although they have the advantage of leapfrogging into the digital economy, the productivity divide between skilled and unskilled populations could lead to a bigger gap in income disparities.
For example, China and India, which were once leaders in a global economy, have already embraced the new reality and are set to capture the lion’s share in the future. Yet, I believe, they need to contend with the lower end of the pyramid, which is still not productive enough. Eastern Europe, Africa and Latin America will be impacted in the same way.
I believe the theory of the population dividend is being turned upside down with the advent of the new technologies. There will be a higher spend due to an increase in productivity and as a result, higher consumerism, but productivity imbalance and earning disparity could cause some social tension.
According to many studies, and this is more a guess, at least 30% to 40% of traditional work is expected to be displaced. This is huge when taken in the context of the countries I talked about. How to manage this huge population divide, where developing countries have leapfrogged into the digital economics, thereby creating great productivity, but at the same time displacing the bottom end of the pyramid, is what we’re going to see in the next couple of decades.
Knowledge@Wharton: Which industries do you expect to be disrupted the most, and which ones will be disrupted least, and why?
Karra: Banking, followed by automotive and manufacturing, will be disrupted the most. There are disruptions in medical technology, for example in diagnostics, gene mapping, drug formulation and modelling, including simulated clinical trials, but they still need to reach the critical mass of industrialisation. Therefore, I believe health care will be disrupted the least, at least for the next 20 years. Likewise, teaching will remain, though the mode of delivery will become more digital. My sense is that jobs in agriculture, natural sciences and conservation are going to be big.
Knowledge@Wharton: What do you think will be the impact on the kinds of skills and knowledge that will be required for the future? Your idea of learnability becomes quite relevant here.
Karra: Historically, we have seen that waves of automation did not destroy jobs completely, but they did created adjacencies and thereby created high-value skills or high-value tasks. I feel it’s critical for the governments and the corporates to map the adjacent skills around these areas, whether AI or automation, and rotate people who have the skill to do these jobs. Therefore, the key skills required in the future will be less technical and more around the learning quotient as I call it, of an individual. This is nothing but the ability to unlearn and acquire new skills. The ability to adapt new mental models and applications will be the key to individual success. So, it’s less about technology and more about ability to learn the adjacencies and the higher value-adding skills of these jobs and tasks.
Knowledge@Wharton: Can you give an example of how a job might change to incorporate more of the learning quotient as you describe it?
Karra: Let’s take testing in a software job. Traditionally, most of the software testing has been very manual in nature. Once the software is developed, testers are deployed to test whether the software is working or not. These guys are literally like hackers. They get into the software and try to destroy its functionality to see how solid it is. But testing is being completely automated.
So what does a manual tester do? He looks at the adjacency of that particular testing skill and looks at how to create various models in which the automation can be tested. He can create frameworks or some kind of functional models where the testing methodology can be used, but he himself is not a tester. He is more a creator of these large situations where the automation of the testing can be conducted. This is a perfect example I can give you about adjacency of skills.
Knowledge@Wharton: It’s a big challenge when job requirements change so drastically. What should be done to bring about a harmonious transition for people who lack the skills that will be needed?
Karra: This is a question which even I am struggling with. I personally think that those who are displaced and equally show the disability to acquire new skills or to reskill, and therefore cannot stay relevant, will be a big social liability. And that’s going to be a big task for governments and companies, equally.
There are three things that I can think of. Let’s take the low-skills jobs — a cleaner, an elevator operator, a waiter, or a security guard. These tasks are clearly prone to automation, but they could be viewed compassionately. The onus is on individuals and corporates to not necessarily automate these jobs as quickly as they could. The other ways are social security and universal basic income. These could mitigate the individual loss to a certain extent, but I doubt if they would succeed in restoring human dignity and self-respect.
The desire to be productive is innate to human nature. Any disruption to this human right could be counterproductive. I would move more towards being compassionate to some of the lower-end jobs which do not necessarily give you quantum productivity, but at the same time there is no necessity for automation. Going a little slow on them is one way to look at the whole thing, apart from other things like social security and universal basic income which could be important, but they’re not necessarily the only answer.
Knowledge@Wharton: Those are great points, Srikanth. The challenge, I think, is that jobs that will be automated are not only the low-income jobs, but also many high-income jobs. One example that I keep hearing over and over again is in the medical field. For instance, if the ability of complex AI systems and powerful computers to diagnose what patients are suffering from is greater than the human ability, then what does the role of the doctor become in the future? And what can be done about that?
Karra: We did talk about how gene mapping, clinical trials and drug testing is being automated. Radiology definitely has reached great proportions from x-ray machines to ECG machines to the CT scans that we have today. There’s a phenomenal amount of diagnostic capabilities with AI and deep learning added on to it. At the same time, when it comes to high skills there is more possibility of adjacencies.
Let me simplify it. In a cleaner’s job, which can be automated by robotic vacuum cleaners, there cannot be an adjacency because cleaning is cleaning. But in the case of high-end skills like those of a medical doctor, there still is a requirement for the person to have some kind of expertise or develop adjacency skills.
For instance, even if you have a CT scan report which has automatically diagnosed a particular disease, there is still a need for a human touch, the human element, to humanly explain to the patient what these symptoms are, what are the outcomes of the diagnosis, and what are the various possibilities right from medication to natural healing to diet-related solutions.
These are interpretations which a doctor can still do. So, he can develop more adjacencies which go beyond just diagnostics, as compared to the low-end skills. As I said earlier, health care will be least disrupted as compared to banking or other industries like automotive.
Knowledge@Wharton: If there are going to be these kinds of major disruptions in both high-income jobs as well as other kinds of jobs, what do you think will be the impact on incomes and wages? What kind of work do you think will be compensated most highly and what will be most vulnerable to wage pressure?
Karra: There is going to be, no doubt, an income disparity. High-end design and architecture, technical skills will be compensated highly. They will be at a premium. My favorite quote is: “A music composer will be paid more than a music producer” because music itself is getting disrupted by digital technologies. The demand for these skills, and as a result premium for them, will be much higher. I think managerial jobs and jobs which are repetitive, including the new adjacencies of the low-end jobs, will be under wage pressure. The adjacencies of high-skill jobs, like in the doctor’s case that we talked about earlier, could still command a premium. So, for the low-end jobs, while there could be some adjacencies, they’d be at a lesser premium than the high-end, technical, architectural, and design jobs.
Knowledge@Wharton: If the idea is to emphasise things like creativity, and that’s the kind of work that will be most highly compensated, how will the hiring process have to change? How will you assess whether somebody has a high learnability quotient or not?
Karra: In simple terms, the velocity with which one acquires most skills and knowledge in the least possible time can be defined as the learning quotient. In order to assess this, you need to understand the candidate holistically and not just through the interview methodology. It could be through tests or through what he’s done in the past few years, the experience that he has brought to the table, his ability to adapt to different skills and how he applies them in various situations. So, it’s a situational kind of interviewing rather than just a pure technology-based interview or the traditional interviewing methodology. You can gauge a person as to what kind of adaptability he brought to the job, how many skills has he acquired in the least possible time, how well could he adapt to adjacent skills.
For instance, in the case of software testing, did he understand the whole testing methodology and how quickly did he adapt it to automated testing? What kind of test cases could he produce at a fast pace? These are nothing but the history of adaptability, the ability of a person to relearn and unlearn. I think these are the measures that we have to put in place. We have some case studies around this. I’m sure many people will come up with their own ways of assessing the learning quotient.
Knowledge@Wharton: Based on everything you’ve said so far, what do you think are the implications for companies and for educational institutions? How will they have to change?
Karra: Almost all companies are paying more attention to reskilling and have also started investing more in learning. But educational institutions, barring a few, have not yet grasped the new methods of learning and developing the curriculum for the new-age skills. Some of the universities that talk to us are still using the old, traditional technologies and not able to grasp the content and the curriculum around which the new-age skills are getting evolved. But I think it’s better late than never for industry and educational institutions to partner more closely than before to bridge the gap.
Knowledge@Wharton: What do you think should be done today to prepare for all these major changes that are coming down the road?
Karra: We overestimate disruption in the short-run and underestimate it in the long-run. The truth is somewhere in between. Just to give an example, the advances in cloud started in the early 2000s. However, it is yet to catch up after two decades. But it is expected to reach exponential growth henceforth. So catching the trends early on, and more importantly, mapping those skills and knowledge to the best possible extent, to the greatest detail, creating learning roadmaps and content around it, and most importantly, creating a learning ecosystem and culture both in companies and countries are key. I don’t think it would suffice to just measure performance of an individual or a company, or for that matter, even a country. Measuring capabilities and learning quotients hold the key.