Tuesday 24 November 2015
A recent report from Merrill Lynch presents a vision of the future where robots perform up to a third of all jobs and particularly those roles that are low-paying, manual or high-risk.
However this is a vision of dystopia with massive social divisions that will be familiar to most readers of classic science fiction. The human labour being replaced by robots will result in many people being unable to earn a living wage while a select few will reap the rewards of this ever present and relentless technology. Deloitte and the University of Oxford have also separately predicted that one-third of current jobs will be automated or replaced by robots in the next 20 years. A further, more direct, example of this continuing change in the jobs landscape is Google’s recent release of their code for the Artificial Intelligence engine - Tenderflow. The software currently drives Google’s image search and translation services. Public access to this code enables a wider group of developers to take one more step towards the automation of tasks that were once considered the exclusive domain of skilled labour.
This change should not come as a surprise. The past 20 years of change offers significant lessons for the next 20. Jobs that no longer exist are telling of the range of methods that have been applied again and again to different roles in the name of efficiency, convenience or simply the attitude that “near enough is good enough”. The early demise of the typewriter pool and the manual switchboard in large offices and the adoption of welding robots in car manufacture was the beginning of the continuous shift away from jobs based around repetitive tasks. Changes to the jobs landscape have not always been a matter of machines becoming direct replacements for humans. A telling example of this change to jobs was the shift from the need for trained typesetters to amateur desktop publishing. Without typesetters, any semblance of good typographic design was forsaken in preference for quick – and dirty – output. This output was also cheaper because it moved the task of setting documents from the realm of specialists to that of generalists. Technology has also enabled new ways of working that move repetitive tasks onto a global and more competitive market. Distributed services such as Amazon Mechanical Turk and GetaCoder allow simple internal business processes to be placed onto a global marketplace for completion offshore by workers in weaker economies who are able to offer lower bids for the work.
In 1997 Evans and Wurster offered “Strategy and the new Economics of Information” in the Harvard Business Review. Even though the work is nearly 20 years old, the lessons of the examples cited in their work remain relevant and document the earliest stages of a process that continues today. Their core assertion that all businesses are information businesses not only remains true but it is now fully proven. The shifting economy of encyclopaedia production is an indicative example used by Evans and Wurster. The production of Britannia was a complex, lengthy and labour intensive work that could not withstand the introduction of the CD-Rom based Encarta. Removing the cost and time taken to physically print and distribute an encyclopaedia was enough to change the prevailing business model. Encarta itself then subsequently failed as a business model when the traditionally labour intensive editing process was removed from the encyclopaedia cost model through Wikipedia’s crowd-sourcing of collective expertise and knowledge. Evans and Wurster’s conclusions reveal that the continuing threat to all existing jobs (and business models) comes from their ability to become unbundled or de-constructed.
The more recent success stories of Uber and AirBnb show that de-constructing existing business models still works and reveals the consequent impact on jobs. In both examples the new business model separates out the individual activities and uses software for what can be automated – in this case matching people seeking a service with suppliers through the mechanism of an app. Both systems then rely upon generalists rather than specialists to complete the service. Uber drivers are regular examples of drivers who rely on GPS services rather than their ‘knowledge’ to reach an unknown destination – evidenced by double parking, erratic signalling and unexpected U-turns. AirBnb relies on property owners – who could be anyone – to provide the expected range of hotel front of house services to guests. Both of these examples of deconstruction reveal the “near enough” attitude to new business models when the attractions of lower prices sacrifice high levels of safety.
Despite the apparently relentless march of technology the situation is not entirely negative. The claim of the Deloitte report hangs on one important phrase, “current jobs”. In the age of robots there are still many jobs that continue to resist deconstruction and automation. It is a cliché but the observation “it is more an art than a science” hallmarks the essential qualities of these still robot-free roles. “Art” is intended to be taken broadly in this sense and is about those roles that involve decision-making based on incomplete information, judgement and ultimately intuition. New roles will continue to appear, develop and evolve in the era of de-constructed business models. This will involve work that requires and ultimately exploits the human capacity to function, make decisions and generally navigate through new forms of incomplete, imperfect and fuzzy information.
By Dr Gordon Fletcher, Acting Dean of the Salford Business School.