What if you need to be an expert by the time you graduate from school simply because autonomous machines and automation have taken over work that is predictable and regular?
The Perfect Storm of Automation, Autonomous Machines and Artificial Intelligence
Dr. Michael Choy | Dioworksgroup
The Perfect Storm of Automation, Autonomous Machines and Artificial Intelligence
Dr. Michael Choy | Dioworks Learning
Director, Dioworks Learning
Adult Educator Mentor, Institute for Adult Learning
2-Time Winner of National InnovPlus Competition, 2017 & 2018
What if going to school means learning from robots and making stuff which gets paid by the government? What if you get a chance to jump grades and do cooler stuff simply because you are already at that stage of professional knowledge and competency? What if graduating from University or Polytechnic is not good enough? What if you need to be an expert by the time you graduate from school simply because autonomous machines and automation have taken over work that is predictable and regular? What if you don’t have the time and opportunity to formulate and develop your expertise? What should you do?
I was in AI 30 years ago and by that, I mean Armoured Infantry, during my national service days. I understood what it meant to fight tanks with tanks. It was really no use using rifles to shoot at tanks. Similarly, we need to use AI to compete with AI because any other way is likely to be futile. To be clear, it is still early days yet for AI but the potential of AI represents a huge leap in the processing capability of machines, moving into areas such as visual recognition, and analysis of big data.
“Fight tanks with tanks, AI with AI”
We need to use AI to learn faster, better and more effectively. Human learning has to increase at an exponential rate because machine learning outscores and outlasts human learning on almost all counts.
How to Better Equip Ourselves in the Face of Artificial Intelligence and Automation
In the story of two men foraging for food in a forest and a bear appearing out of nowhere, we were told that to survive, one does not need to run faster than the bear but to outrun the other person.
In this analogy, if we take artificial intelligence (AI) and automation as the bear, and ourselves as the men who need to outrun each other or if possible, the bear itself, the challenges and our corresponding plans may become clearer to us. Hence, one solution that we may consider is simply to outcompete the other person and not necessarily the bear.
“But how does this analogy work?”
Technology has moulded and shaped the workforce over the years. From improving efficiency to workplace safety, it has forever changed how we approach and do work. In fact, every industrial revolution of the past were sparked by a new breakthrough in technology: the first was brought forth by the invention of steam engines, liberating us from overdependence on animal labour, the second was made possible by using electricity to drive mass production, which helped feed the rising demand of the global population and more recently, the third, where digital technology has helped to connect billions online.
It does not stop there, though. A fourth industrial revolution is on the horizon and it mainly revolves around Artificial intelligence (AI). With improvements in technology, machine learning algorithms are becoming ever versatile and AI can be eventually programmed to perform activities that follow a predictable pattern. These include cooking, cleaning, driving, reviewing MRI scans and many other behaviours. The implication is that many processes performed by human specialists can and will be automated. In fact, in their (2017) report, McKinsey & Co. predicts 60% of all jobs have at least 30% of activities which are able to be automated and this will result in a loss of 400 to 800 million jobs by the year 2030. For companies, this presents higher efficiency and profit margins as AI and automation eliminate human error and limitations. However, for the employee, the outlook turns negative. The encroaching AI and automation (the ‘Bear’) results in stiffer competition as both human and machine race for that job role.
The Global Status Report on Road Safety (2018), published by WHO in December 2018, highlights that the number of annual road traffic deaths has reached 1.35 million fatalities. Road traffic injuries are now the leading killer of people aged 5-29 years globally and the majority of those killed are the pedestrians, cyclists and motorcyclists.
Research by Ohio University (2017) indicates that with driverless vehicles, also known as Grade 5 Autonomous Vehicles (AVs), we can benefit from AI in at least three ways:
- Reduced rate of accidents due to human misjudgement and lapses in decision-making
- Reduced travelling time due to more efficient pathing and avoiding of congestion
- Reduced carbon emission due to more efficient braking and acceleration
Where do humans figure in this equation?
It seems that human intervention is still required whether remotely or physically when the AV is in operation. Currently, most AVs’ autopilot is not extremely reliable and humans still has to take over the wheel when needed. Refer to the diagram on the 6 stages of automation in vehicles.
The 6 stages of automation in road vehicles
In the long run, if full automation were to take place, human drivers will monitor and control the AVs remotely from a control centre, overriding the automation when necessary (e.g. Calderone (n.d.) on Robotics Tomorrow). Imagine a whole army of driving controllers in every state and city, overseeing millions of AVs globally.
Hence, while AI encroaches into the driving industry, displacing millions of jobs, there are new jobs which will appear. Some of these jobs will have better compensation packages than current jobs. The critical question is which of the displaced drivers will be able to take on these new jobs.
Job mobility is important as it can provide you with more career pathways
If you are currently a professional driver, will you be able to outcompete other drivers to take on these new roles? Interestingly, the concept of feeder role is not new (e.g. Career Framework by Mercer). What this means is that employers will require staff with new capabilities. Often, these new capabilities require the layering on of new competencies and mindsets over current professional beliefs and values.
As 2-time winners of the national InnovPlus Competition (2017 and 2018 editions) in Singapore, our team developed chatbots to drive a new experience for online learning. With chatbots for learning, curriculum designers no longer work on lesson plans or storyboards but on conversational trees. It utilises a branching methodology to drive learning scenarios and decision-making. The new designer capability is demanding. It layers new engagement pedagogy over foundational adult education beliefs and principles. Conversational tree designers are required to have strong foundational skills in designing online courses before they can be considered to take on this new role. What makes a potential conversational tree designer? The feeder capabilities include a sensitivity to learner needs, scripting fluency and instructional design.
In the 2019 survey by the Ministry of Manpower (MoM), the highest number of job vacancies was for trainers and assessors, over and above jobs such as accounting and computing.
Going forward, the Continuing Education and Training sector will require even more professionals especially those with technology-enabled learning skillsets.
Based on research published by Carl Benedikt and Michael A. Osborne from Oxford University (2013), jobs requiring:
Automation will not be able to replace the entire workforce, so it is important to understand how AI and us will work together in the future. Homing in the evolving skillsets that are valuable to your job description can help you remain competitive as well as offer you opportunities to progress either horizontally (Across industries) or vertically (Within your industry).
Knowing how to make space for AI and automation will be critical for human workforce development so that human workers perform the roles that involve decision-making and pre-empting errors. Remember technology has the capability to multiply impact enormously so any mistake can be horrifyingly catastrophic.
So when the bear appears, we can either…
Move into spaces which AI may not necessarily encroach into at this point in time
As an Adult Educator trainer and mentor, I have coached more than 10,000, teachers, trainers and instructional designers over the past 20 years – some at novice levels but also many at intermediate and expert levels.
More recently, one of my learners working in a semiconductor industry was told to codify his skills so that the organisation can automate his role. John (not his real name) understood his role so well that besides codifying his skillsets, he also highlighted areas which would need a lot more work and could not be automated. John’s boss was so impressed with John’s meticulous work that he kept John to oversee the automation process while his counterpart performing the same role in Switzerland was retrenched. In short, John outdid the competition by getting into the feeder role, which was to manage the on-going and future automation process.
What John provided was a blueprint on how a worker could get into the feeder role during the automation process; he carried out the following tasks:
- Codified key skillsets
- Value added by signalling where machines cannot yet do the job
- Specialised in plugging the skills gaps
- Facilitated the business process transformation
- Operated at expertise level by reviewing the automation process by putting on strategic, innovative and metacognitive lenses
- Maintained a resilient mindset even though he was given advance notice of a possible retrenchment or job redesign
As with John’s organisation, all enterprises seeking to automate will eventually need experts to manage the automation at the operational, supervisory and management levels. These experts work with intelligent machines to automate some of the current manual processes and in the long run, pre-empt errors as well as override the machines in the event of faults and accidents. The Air Ethiopia and Lion Air crashes involving the Boeing Max planes in 2019 taught us a few valuable lessons. One of which must be the re-examination of how humans think in relation to intelligent machines and automation.
Have pilots forgotten how to fly without autopilot?
Due to the high level of trust built up over the years of usage, pilots stop questioning the validity and accuracy of data put forth by machines and the corresponding behaviour of these machines. To actually rise above the curtain of trust and question the machine data and behaviour requires a different manner of training approach.
“With so much automation in the flight deck nowadays … pilots may feel safer psychologically to keep relying on some autopilot functions ...”
- Prof. Terence Fatt (Singapore Management University)
Effectively, pilots are trained to rely on machines and data to make decisions and this can be disastrous as the two air crashes demonstrate. Deciding when to cart over from relying on machines and automation to human decision-making is not a choice to be made lightly, especially under time pressure. Hence, this decision-making to question machines must be part of the training.
One approach that we have adopted over the years was to design errors intentionally into training, to develop metacognition and expertise. Through errors, practitioners are facilitated to question underpinning beliefs and assumptions during training, thus building elements of metacognition and deep expertise. However, most training in Singapore and globally does not deal with expertise training, only competency-level training.
With AI taking over competency-level tasks, we will short-change our learners if we do not shift learning designs to developing expertise. The race has changed. As AI is trained through massive data and feedback, any behaviour that can be tracked with the outcomes measured and feedback given, AI can and will improve at breakneck speed. Hence, competency-based tasks are perfect for AI to tap on, primarily because all competency-based training (CBT) are underpinned by measurable outcomes. If humans continue to depend on CBT, we will likely be outrun and out-thought over time.
In short, we can no longer win a 100-metre race; we have to stretch it to 110 metres. The intelligent machines can complete 100m in less than 1s (as an example). Humans take on average, about 10s. However, most intelligent machines cannot yet reach 110m. They do not quite have the metacognition, innovative mindsets, ability to pre-empt errors and question assumptions. Humans can complete the additional 10m run only if they are appropriately trained.
However, the majority of the workforce are trained to be only competent. Moving them into the expertise zone (i.e. the additional 10 m) will take a concerted effort to coach and mentor them, so that they can outcompete the intelligent machines and human resources in other countries.
The bears are coming!
We cannot assume things will remain the same. However, if you ask any man on the street how he is preparing for AI, the response is usually a puzzled look or to take each day as it comes. Some expect the government to resolve the challenges brought about by AI. Not many are thinking about the future, nor can they do anything about it. The learned helplessness can be incapacitating and disempowering.
We need a more ingenious and simple approach to get people into feeder roles, by training them to reach expertise levels. The competition will be stiff, the learning will be unprecedented but the opportunities and rewards can be immense and satisfying, only if we can ride the oncoming AI wave – to run and keep on running.
Dr Michael Choy is Director of Dioworks Learning, a technology-enabled learning company with offices in Singapore and Vietnam. They are also the local strategic partner to Udemy, the largest marketplace for online professional courses, in Singapore and Indonesia. As an Adult Educator Mentor, Michael has coached and trained more than 10,000 educators in Asia over the past 20 years and produced numerous online courses including one on Designing Errors for Learning and Teaching (DELETE™). His current focus is to produce chatbots for learning.
Calderone, L. (n.d.). Autonomous Cars – Safety and Traffic Regulations. Accessed on May 9, 2019: https://www.roboticstomorrow.com/article/2018/10/autonomous-cars-–-safety-and-traffic-regulations-/12722
Mercer LLC (2017). Building for an Unknown Future: Leap forward with a Career Framework. Accessed on May 9, 2019: https://www.uk.mercer.com/content/ dam/mercer/attachments/europe/uk/uk-2018-building-for-an-unknown-future-leap-forward-with-a-career-framework.pdf
The content & opinions in this article are the author’s and do not necessarily represent the views of RoboticsTomorrow
This post does not have any comments. Be the first to leave a comment below.
Post A Comment
You must be logged in before you can post a comment. Login now.