.png){kind=logo}
A while back, well before the onset of COVID-19, child psychologist Alice Gropnik went to the Defense Advanced Research Projects Agency (DARPA) government research lab to talk about babies.
This might seem like a strange place for a child psychologist to visit, especially if not somehow affiliated with the project, which she wasn’t. But scientists there were interested in exploring the idea that the way we learn as infants or toddlers might be applied to help artificial intelligence (AI) algorithms become even smarter.
It’s a novel concept. Indeed, some would argue the idea that one of the most revolutionary technologies of our time could actually benefit by patterning itself after human beings who haven’t yet put a complex sentence together might even be called “radical.” But the breadth of learning ability, also known as plasticity, that babies show when consuming, processing, and learning from small amounts of data - is something that even strong AI tends to struggle with.
When babies play, they learn. And they do that in ways unlike how AIs learn. When a child lets go of a toy and picks it up then lets it go again, it’s doing so because it is curious about the rules of dropping and retrieving objects. In the process of that seemingly mindless process, it is actually learning about touch, gravity, rising, falling, velocity and time as well as emotional responses that accompany losing then retrieving something. If you were to take the toy away and replace it with, say, a block, the child might resume the same test to see if this different object triggers varied results. Knowing absolutely nothing about scientific method, they’ve gone through the steps of forming a hypothesis and testing to reach a conclusion.
As Gropnik put it in a Wall Street Journal essay, “one of the secrets of children’s learning is that they construct models or theories of the world,” while entertaining themselves. AI, however, struggles with these types of processes. They’re great at many things but bog down when asked to learn the same way all of us did as children. They need specific data and examples to thrive.
In many ways, we as adults need to be more like babies and less like robots. As we age, we too get bogged down with our preconceived notions of the reality of the world. We often lose the ability to think more conceptually. To experiment. To constantly reassess or even challenge our views of the universe. And this inability to modify how we process information and adapt could end up costing many of us income and jobs before long.
According to the World Economic Forum (WEF), 133 million new roles are likely to be generated worldwide by next year as a result of technology-generated divisions between humans, machines, and algorithms. Even as automation nudges some people with rote, repetitive, or redundant jobs out of work, the thinking goes this Fourth Industrial Revolution will generate millions of new job opportunities.
The trouble is: there aren’t enough people possessing the necessary technical skills to staff all those jobs. The WEF, for instance, estimates more than half (54%) of all employees will require significant reskilling by 2022 to fill those positions. But if we, as adults, cannot learn effectively, cannot get back to the plasticity we exercised as children - we might not have the place we want, as individuals or humans, in the future of work.
To remedy this, people and organizations must refocus personal and corporate learning on cognitive skills and foster a culture of growth mindset that will enable them to roll with the inevitable technological punches that will come their way.
What’s that mean? Well, the obvious answer is that organizations need to be willing to retrain employees. And employees need to recognize they could become obsolete by sitting on particular skillsets; they have to keep learning.
It will be critical for people to go beyond their comfort zones and really understand – at a fundamental level – how the technology coming their way works. For example, Gartner predicts nearly 70% of routine work currently done by managers will be fully automated by 2024. They’re talking about AI taking over repetitive ‘mindless’ jobs so people can spend less time managing transactions and more on learning, performance management, and goal setting. But for AI to do its job properly, people still need to know how to work with it. To input the right data and instructions so the AI can process, analyze and provide viable recommendations for necessary actions.
Think about it this way: when first learning to use a computer today, we aren’t required to know how to code. We don’t need to understand the inner workings of motherboards, semiconductors or random-access memory. We do, however, need to understand the basic differences between hardware, software and cloud applications. We also need to know which office suite applications are relevant to us and double down on learning those that are instrumental for doing our jobs. As new features come along, we must be sure to check those out too because they could greatly assist us in being more productive, efficient and successful at our job.
Similarly, employees in the future of work will need a basic understanding of how the algorithms behind AI work. What are the basic assumptions that drive their calculations? What are they optimizing for? How do they process signals for success, and how might those signals be manipulated? Those of us who are mathematically or technologically challenged might feel a little intimidated by this responsibility. But nobody is asking every employee to become a data scientist. Rather, they’ll just need to understand what makes AI tick and do their best to input good, clean, current data and instructions to achieve the most beneficial results.
Employees will also need to stretch their brains and gain exposure to broader technological trends likely to drive future business. We live in computational times that will underlie almost everything. Having some exposure to what will drive the future of work will be as mandatory to success as reading, writing, and arithmetic were for all of us in elementary school.
In its Future Work Skills 2020 report, the Institute for the Future (ITF) calls out a few useful skills that could be considered essential in the future of work. The ITF defines these as:
These 10 skills from the ITF could form the underpinnings of any training program for the future of work. What all have in common is that they are new types of cognitive capabilities. While not evidently in demand today everywhere, they are going to become prerequisites for the not-so-distant tomorrow. By not having these key proficiencies, people and organizations risk losing valuable human talent which simply has not been giving the chance to success in the future of work.
Individuals should not wait for employers to mandate change. By returning to who we were as inquisitive kids, and committing to life-long learning, we can be a valuable part of cognitive automation systems built on AI.
Get social with us!
Follow us on LinkedIn: Cognitive Automation Community
Follow us on Twitter: CognitiveAutomation
.png)
