Site icon Future Of Work

Leveraging Our Comparative Advantages Over Robots: The Bottlenecks of Automation

While waiting to board an inter-city bus in Montréal, Canada, an interesting conversation on the bottlenecks of automation was struck between a mid-aged American bus delivery driver and me.

Wallette faces the risk of losing her job to a robot in the coming years. The long-haul transport  truck driver of school buses isn’t worried, even though her job faces a 92% chance of being automated in the years to come.

Stopped at a Montréal inter-city bus station. Wallette is focused on the 3000-kilometre trip ahead of her. Her rig is loaded with brand new yellow school buses destined for Denver, in the United States.

She does not foresee a day when robots will be behind the wheel of those vehicles, either. “Human bus drivers will never be replaced by self driving buses”, she says. “There are [sic] just too much coordination, agility and random chances and changes” involved in getting school children to school and back”.

And she has a point. Would parents be comfortable waving goodbye to their children at the bus stop with a robot in the driver’s seat?

Wallette’s observation is a great example of one of the “bottlenecks of automation” – where humans will continue to have an edge over machines in different jobs and job tasks.


The Race to Build Autonomous Vehicles

Ride-sharing start-ups, tech giants, and automakers are racing to build fully driverless (or autonomous) vehicles. Uber’s self-driving Otto trucks were unveiled at the end of 2016. Otto’s system differs from its next closest competitor – Tesla’s Autopilot – as it offers true ‘Level 4’ autonomy. Chinese tech giant Baidu says it will have the same technology by 2019, while mobileye and Delphi promise to make Level 5 (completely without human intervention) self-driving cars by the same year.

Driverless cars (and to a lesser extent drones and airships) are disruptive technologies that will transform the world of transportation and logistics. They are already challenging how we move goods, services and people – as well as the business models, governance structures, safety and ethics protocols, and jobs that are traditionally associated with them. These concerns have forced the US Department of Transportation and the White House to issue an initial road safety guidance on automated vehicles at the end of 2016.

Today, trucking is the biggest job sector in 29 American states. However, the McKinsey Global Institute’s latest research reveals that jobs in the industrial trucks and tractor operators sector have a 92% automation potential rate. This means over the next decade, robots and artificial intelligence could replace up to 1.7 million American middle-class trucking jobs and another 1.7 million taxi drivers, as reported by the Los Angeles Times.


Table 1: Automation Potential and Wages for US Jobs | McKinsey Global Institute

Click the above photo to access the interactive table

There is also another important distinction to note. Wallette specified that “school bus drivers will never be replaced by self-driving ones” and not “industrial trucks and tractor operators”. School bus drivers interact with children on a daily basis and this requires originality, service orientation, manual dexterity and gross body coordination.

“How would a robot know if a student will be sick and not attend school or going to be two minutes late to catch the school bus because of blackout?” she shared. “There are just too many random chances and changes” each day, for each household, for each student riding the school bus. You would need good old human judgement for that – something that we (human workers) often take for granted.     

According to Michael Osborne, Associate Professor in Machine Learning at Oxford and the Co-Director for the Oxford Martin Programme on Technology and Employment’s 2016 report, automation is making some jobs and job tasks go extinct, but not all work skills or tasks are equally replaceable. There are limits or potential engineering bottlenecks to automation. In short, robots and intelligent machines (currently) are not so good at things that make us human — such as caring for others, building relationships, making complex judgements, and coordinating our bodies, and making sense of sight, sound, touch, taste, emotions, and pain.

Osborne identified 1) creative intelligence, 2) social intelligence, and the 3) perception and manipulation of the world around us as the top three potential engineering bottlenecks of automation. This infographic explains these three bottlenecks – and innovation that is trying to move into these new technological frontiers.

Infographic 1: Our Comparative Advantages: Top Three Engineering Bottlenecks of Automation


The Human Element: Leveraging Our Comparative Advantages

These three engineering bottlenecks can provide us insights into what human-centred automation and near-future human-machine complementarity could look like, creating a new division of labour between human workers and machines.

A recent NYT article entitled “Learning to Love Our Robot Co-Workers” argued that one growing area of automation is the ability for robots to collaborate with their human counterparts – in short – collaborative robots, or “cobots”. For now, this “collaborative ability” has allowed the e-commerce and logistic giant Amazon to increase its robot workforce by 50% from 30,000 to 45,000 in 2016 and it plans to increase its total number of full-time human workers in the US by around 50% (or over 100,000), according to the company’s Q4 2016 earnings report. However, it is still unclear how autonomous vehicles could collaborate or co-work with their human truck drivers.

As mentioned, trucking is the biggest job sector in more than half of all American states. Therefore, transitioning workers away from this sector with retraining and up-skilling schemes that are human centred will be critical. In addition to innovation in new forms of social protection floors (to buffer the socio-economic and livelihood stresses and shocks of unemployment), we will also need to rethink our education and skills development systems. The three engineering bottlenecks identified by Oxford give us glimpses of emerging areas of focus for education, training and skills on an individual level, as well as future workforce strategies. “Change won’t wait for us: business leaders, educators and governments all need to be proactive in up-skilling and retraining people so everyone can benefit from the Fourth Industrial Revolution,” says the World Economic Forum.

Finally, according to the World Economic Forum’s Global Agenda Council, the top skills in 2020 are those that smart machines and robots are not so good at (yet), such as complex problem solving, critical thinking, creativity, and people management, just to name a few. These top future skills are in areas where humans have the upper hand vis-a-vis the bottlenecks of automation. In addition to up-skilling you and me, the challenge question of digital inclusion and technological literacy, trying to address how we ensure no one is left behind in the digital age must be at the forefront of policies and strategies.


Table 2: Top 10 Skills | World Economic Forum


Tell us what you think about the engineering bottlenecks of automation and how we can better prepare ourselves for a changing world of work in the comment section below.

You can also download our Infographic summarizing the three top bottlenecks of automation by clicking on the


Further Resources:

Featured image is from Flickr by Lwp Kommunikáció. 

Exit mobile version