Published on the 18/10/2016 | Written by Anant Gupta
Think of the future of automation. What do you visualise, asks Anant Gupta…
Humanoids in our offices? Robots running our factories? Driverless cars on the roads? Watson, Deep Blue, AlphaGo? Or rather, the future of automation is here and now, just it is as far from an edge of the seat sci-fi thriller that you might imagine. Automation is no longer an option in the 21st century enterprise. Across industries, it is already driving efficiency, productivity, agility, adaptability and optimisation. Yes, there is growing apprehension about automation devouring jobs. But this fear is not unexpected if we take history as our guide. The Industrial Revolution was met with what John Maynard Keynes called “technological unemployment” as a “temporary phase of maladjustment.” More recently, the Knowledge Revolution, the internet and cloud computing were met with such concerns. Each of these chapters was finally turned as a net positive. And today, we stand at a similar crossroads with the automation revolution. Automation – or autonomics, including both automation and AI – presents perhaps the greatest opportunity the world has seen in a long time. An opportunity to unlock the power of technology to break through the global economic logjam. However, in order to realise its potential, we have to get past our apprehension and recognise three basic ground rules of automation: It is a partnership As the Open letter on the Digital Economy pointed out, the idea of robots eating away jobs “assumes that we are powerless to alter or shape the effects of technological change on labour.” Sure, the power of AI is awesome. But it is nothing compared with the power of a partnership between humans and machines. A small case in point being the power of the cyborgs in free style chess – a combination of man or machine that has proven to be far superior than either. This partnership is reflected in the increasing switch-over to the new, more powerful AI – augmented intelligence. It automates activities, not jobs A recent article in the McKinsey Quarterly analysed workplace automation to find that while around 60 percent of occupations could have 30 percent or more of their constituent activities automated, fewer than five percent of occupations can be entirely automated using current technology. This led the authors to conclude that “Very few occupations will be automated in their entirety in the near or medium term.” On the other hand, they pointed out that the automation of certain activities would necessitate redefinition of jobs and transformation of entire business processes. In other words, the possibility to automate certain activities, opens up vast opportunities for augmentation of roles played by people in the workplace. For instance, if the routine and tedious task of reading and updating the system with information on ongoing new developments could be automated at, say a law office, a hospital, a media outfit or any knowledge-based industry, it could free out the time of people working there to do far more productive, creative and beneficial work. It is a (well-planned) journey And at every stage, there needs to be a successful transition – an up-placement – of the workforce. At every stage, there has to be a sharp focus on re-skilling of people, to not just move successfully into the next level, but to effectively fill all the new related jobs that will be created within each scenario. This is not wishful thinking but reality, as discovered in a study at the London School of Economics, which looked at the impact of industrial robots on manufacturing in 17 developed countries. They found no evidence that the robots reduced total employment. While robots did seem to replace some low-skill jobs, their most important impact was to significantly increase the productivity of the factories, creating new jobs for other workers. The fact is that technology has been a lifeline of human progress. Adoption of new technology has quite literally marked key milestones on the timeline of development, therefore the AI and autonomics revolution can be viewed with great optimism and confidence. A recent session on ‘Artificial Intelligence: Friend or Foe?’ at the Global Conference by The Milken Institute had among the panelists Shivon Zilis, Partner and Founding Member of Bloomberg Beta. He said that as long as we kept a sharp focus on using machine learning for good, providing the right balance of power, and the right standards and regulation, the benefits of AI could have a transformative impact on our future. And Professor Stuart Russell of University of California Berkley, cautioned that we must first “envision a future” to realise the true potential of AI, and then “make a transition plan to that destination.” At its very foundation, the raison d’être of technology is to trigger and harness disruption for the benefit of people in a collaborative process. Automation is a new frontier of this human-machine partnership; one that will be good for business, good for people and great for society as a whole. Anant Gupta is president and CEO of HCL Technologies
Look back at any major technological development; in each case, its value emerges from a collaboration between people and technology. It’s not an ‘either or’ scenario. We are the architects. We are the users. And, without a doubt, we are the beneficiaries of an automation-enabled tomorrow.
There is only one way to make automation work right. And that is the pro-human way. We cannot jump in blindly. Automation is a pragmatic journey. It must be adopted in waves, based on the maturity of the prevalent IT estate within the enterprise. This requires a deep assessment and a thorough mapping of these technologies; before even attempting to apply automation levers. And once you do so, it must be in sustainable layers, and in partnership with people. The moment you take a methodical and analytical approach to applying autonomics – you can start looking beyond the hype and apprehension; and start looking at making positive impact with real, repeatable and consistent outcomes.
It is amazing how short-sighted people are when they look at the world through an IT lens. They see it as black or white, on or off. The real world is not like this. It is donkey-deep in nuance and this can result in some technologies evolving slowly… like speech and automation.
Back in 1997 Bill Gates predicted speech would replace keyboards. It should have happened well before now. Why not? …because of the idiosyncrasies of the human environment. When two people converse they adjust their interpretation based on age, accent, environment, background, and a whole lot more. We can change what someone says in our own mind based on these factors. For Donald Trump it’s 90%. Computers will have an IQ of 30 until they are as aware of their environment, and like a human speaker can factor in a speaker’s background and other nuances. Shall we say 2100?
Automation is in the same class. Robots work in factories because their functions are highly repetitive and they don’t have to deal with exceptions – events they are not programmed to handle. Until we can say the same in a non-factory environment automation will be a non-starter.
The problem is you can’t automate what you don’t know. The inability of business to specify precisely how people work has dogged IT since day 1. It killed the BPMS market and after 50 years the IT system failure rate is higher than ever.
Technology enables us to introduce new business models like Uber and Amazon, but these purely use conventional IT techniques in a more productive process than incumbents. The type of automation that can put people out of work is a long way off. Uberization, on the other hand is just around the corner.