28 November 2014
This article first appeared on E&T Magazine
PA’s Rob Gear, an IT expert and futurist at PA Consulting Group, discusses the emergence of robots in place of human employment.
The employment market is used to self-managing production lines, but is relatively unprepared for these robots to take on the knowledge-based tasks commonly regarded as the preserve of human supervisors, according to Rob Gear, Futurist at PA Consulting Group's Foresight and Innovation Unit. He believes that significant swathes of salaried personnel will have to adapt to a “new world” where commerce, policy makers, and education strategists will need to understand the reasons why we need to adapt.
Gear is author of a briefing paper, 'The Robots Are Coming', showcased at a PA Consulting Group's 'Mind versus Machine' innovation event that took place in London earlier this week. He says that we should not assume that, like other technology-based 'revolutions' where the disruptive technologies cause contraction in the short-term that will ultimately fuel economic expansion over time, new employment and new markets for electronic goods and services will eventually evolve.
“In previous waves of automation resulting increases in productivity have – in the end – generated wealth,” he says. Despite concerns similar to those now being voiced, new jobs were created to replace those lost; but there's no evidence that history will repeat itself. “The new wave of automation is different – it has the potential to automate the work of entire occupational areas.”
Cashiers, airline check-in desk clerks, travel agents, and financial traders, for example, could end-up almost completely obviated by customer-facing technologies with ML/AI back-ends. Think about that the next time you self-serve in your local bank or supermarket: could be that within a decade there'll be no in-store staff, just self-serving customers kept in line by batteries of watching surveillance cameras.
Gear's message is not altogether one of concern: “There are still many human skills and qualities that machines cannot easily replicate – those that are not easy to standardise or codify into an algorithm.” They include leadership, motivation, intuition, empathy, abstract reasoning and lateral thinking. The business author Don Peppers has noted that discovering new problems is something that computers can't really do – and are unlikely to be able to do in our lifetimes, Gear says.
Part of the solution, he suggests, is for humans to develop ways to collaborate with computer systems on a more equitable basis: rather than computer systems being our tools, they become proto-colleagues who make autonomous contributions to emerging tasks an challenges. First we have to accept that new opportunities for collaboration with 'machines' could add significant value, because there are many things a machine can do more efficiently than a person.
Explains Gear, “Ultimately, we could achieve a greater degree of 'human-computer symbiosis'. This would mean people and machines co-operate in making decisions and controlling complex situations without inflexible dependence on predetermined programs.”
Q&A: Rob Gear, IT expert and futurist, PA Consulting Group
E&T: What do PA Consulting Group's clients look to a Futurist such as yourself to provide them with? Is your role to an extent facilitating the channelling of human predictive modelling intuitions that exist but are not typically used by enterprise strategists?
RG: It is a common misconception that the role of a Futurist is prediction. I see it more as provocation and in particular helping clients to confront and explore uncertainty and their own assumptions. Within an organisation individuals often hold many different (and sometimes conflicting) views about what the future holds: a futurist will help them to have an open discussion to share and explore these views and arrive at consensus as to what a preferred future would look like, i.e. the 'vision'. The aim is also to create and plan for other plausible alternative scenarios rather than adopting a single linear view of the future that is often based on extrapolation and continuity with the present.
E&T: Baxter could be among the first of a generation of robots designed to be easily taught by humans to perform a range of tasks without complex programming. Many at-risk workers have had to accept the fact that robots on assembly lines, say, can perform bigger jobs faster and better - but they never had to 'teach' them how to do so. What are the chances that we will encounter a scenario where people will decline when asked to 'teach' the coming generation of automated robotic systems how to do their jobs?
RG: Tricky question to answer in general because people’s attitudes and motivations differ dramatically according to geography, culture, socio-economic circumstances, etc... Ultimately there will be some who accept the idea of working collaboratively with machines, and those who feel less comfortable. It will be important for organisations to make the case for the use of the technology and do all they can to coach and support their staff in the new ways of working.
E&T: You say that the skills machines cannot acquire include leadership - yet some future-watchers are suggesting that computer learning/AI systems could be used as decision-support tools that usefully advise company leaders on decision-making and inform other strategic functions. Is this a development that you also foresee, and would you expect senior executives' notorious techno-phobia as a key factor that needs to be addressed in order for this development to prove its worth?
RG: The key thing here is the definition of leadership. Machines are already capable of providing advice based on their analysis of data with some degree of 'intelligence' - but this alone does not constitute leadership. The role of a leader is to guide, inspire, set the ethical and moral tone, and direction of the organisation as well the day-to-day tactical decision making and machines are not able to do these things yet. In the future, it looks probable that we will eventually see Artificial General Intelligences capable of their own independent reasoning and problem-solving; but these will be non-human-like intelligences. We have a tendency to anthropomorphise AI as being human-like - but it will be very different.
E&T: You have said that "Machines can bring improvements in efficiency, make fewer errors" - this is going to prove a moot point on the quest to achieving greater degrees of human-computer symbiosis. In order for human to more equitably collaborate with smart robotic 'colleagues' does there need to be new thinking about how each party’s performance is assessed – e.g., should robot colleagues be subject to the same performance reviews that human co-workers must undergo?
RG: We are only beginning to explore what human-computer symbiosis actually means and how its effectiveness can be measured - but it seems reasonable to me that machines should have their performance assessed. How this is done will depend very much on the kind of work being undertaken – it is very easy to measure the output and performance of a manufacturing robot for example, but perhaps less so to assess the impact of machine-based advice or guidance. As AI’s capabilities increase we will require new ways of evaluating performance compared with that of humans... In the future where we potentially have AI’s that continually optimise and improve themselves towards the achievement of their goals.
For thinking and insights on digital business, click here or contact us now.