Brave new world

When the likes of Stephen Hawking and Bill Gates warn against the development of artificial intelligence, you know we’re in trouble... or are we?

Stephen Hawking, in particular, has made a big call on the topic, suggesting that it could spell “the end of mankind”. At the other end of the spectrum, AI advances are likely to free up humans from more mundane tasks. To do what, though? Be more human?

In the arena of field service, technology has obviously made things easier, but advances have largely been about replacing paper-based tasks with digital solutions. Not that long ago, everyday mobile devices were but a twinkle in inventors’ — eyes and now it’s impossible to imagine life without them.

But how will we feel when mechanised versions of ourselves are roaming the streets, performing tasks we now consider too menial? Will driverless cars change the landscape? Will customers simply print their own spare parts and utilise a robot to replace them, or will they be delivered by drone?

Taking it down a notch, which jobs are most likely to be replaced and how far into the future are we talking? If you’ve only recently entered the workforce, you will most likely see massive change in your lifetime. For those with retirement on the near horizon, probably not so much.

Identifying at-risk occupations

In 2013, researchers Carl Benedikt Frey and Michael A Osborne from the Oxford Martin School at Oxford University published a paper titled ‘The future of employment: how susceptible are jobs to computerisation?’.

Frey and Osborne analysed the impact of computerisation on 702 occupations, utilising a methodology they developed which considered recent advances in machine learning (ML) and mobile robotics (MR). Computerisation itself is classified as ‘job automation by means of computer-controlled equipment’.

The upshot was a ranking of the occupations based on probability. The authors suggest susceptibility is heavily influenced by three factors which they identify as “bottlenecks to computerisation”:

1. Perception and manipulation. Robots still lag behind humans in terms of depth and breadth of perception. Structured spaces such as warehouses, factories and airports are more easily navigated by robots, whereas unstructured spaces, such as homes, represent a more challenging environment as machines “struggle to identify objects and their properties within a cluttered field of view”, according to the authors. Manipulation in turn is influenced by perception capability, and robots also struggle to handle irregular objects. One–nil to humans.

2. Creative intelligence. Frey and Osborne reference Margaret A Boden’s The Creative Mind: Myths and Mechanisms when defining creativity as “the ability to come up with ideas or artefacts that are novel and valuable”. By this definition, ideas comprise concepts, poems, musical compositions, scientific theories, recipes and jokes, whereas artefacts are tangible objects including paintings, sculptures and machinery. The ability to create either requires a deep repository of knowledge and, while some creativity elements are simple to replicate using automation, the stumbling block is the ability to state creative value. The authors put this down to values being variable and suggest that arguments about creativity will stem from disagreements about value... stay with me here. They propose that, even if it were possible to identify and encode human creative values, to enable a computer “to inform and monitor its own activities accordingly, there would still be disagreement about whether the computer appeared to be creative”. It’s a bit like a snake eating its own tail.

3. Social intelligence. This encompasses tasks that involve negotiation, persuasion and care. No chance of these traits being automated? Think again. Whole brain emulation will potentially unlock the secrets behind our innate ability to ‘shorthand’ messages to one another, lessening the need for complete articulation of a thought or response. It’s only a matter of decades away, according to the researchers. Once this happens, it’s anyone’s game — this is probably what Hawking and Gates were referring to when they warned the world.

Frey and Osborne developed a subset of variables that operate under each of these bottlenecks to further inform their research and implemented a range of measures to mitigate bias and to correct potential labelling errors. They ranked each of the occupations from least to most likely to be replaced by computerisation — and it’s good news for first-line supervisors of mechanics, installers and repairers, who rank second only to recreational therapists as least likely to be replaced. Conversely, telemarketers are doomed — no surprises there, as it’s effectively already happened. In fact, based on the outlined bottlenecks, most rankings are as expected.

Entire industries will disappear

Self-proclaimed ‘futurist’ and TEDx speaker Thomas Frey suggests that by 2030, over 2 billion jobs will cease to exist — but this change will give rise to new opportunities. We may see less requirement for drivers, for example, but we’ll need automated traffic designers and architects, as well as emergency crews for when things go wrong. 3D printing output will replace traditional building materials, but printers will need to be serviced by someone.

Frey says highways designated for driverless cars only will be common in 2030, capable of accommodating 10–20 times the number of vehicles they can handle today. He says, “As the transition is made to driverless vehicles, the number of vehicles sold to individuals will begin to decline and a growing percentage will be to large fleet operators offering the new ‘transportation on-demand’ service. In response to declining car sales, the automotive industry will adopt a ‘selling transportation’ model where, rather than selling cars to fleet operators, car companies will begin charging a nominal per-mile charge. Fleet operators will love the arrangement because there will be no large upfront purchase price, but instead, only a small monthly fee based on the number of miles driven.”

Where to next?

We’ve experienced massive change in the past — the Industrial Revolution, for example — but such change served to move people into other jobs. Widespread automation has the potential to lead to high unemployment levels, and economists are grappling with how to handle the shift. Options in the mix include decreasing working hours or implementing a universal basic income scheme, whereby citizens will be paid a uniform amount across the board.

We won’t really know the true impact till it happens, but it’s a fair bet that today’s students of mechatronics engineering will be well placed tomorrow.

Image credit: ©jim/Dollar Photo Club