Author: Andrew Liew Weida
hi my name is Andrew and I want to share with you in the next 1 hour or so is to talk about the AI revolution of human capital. This talk is for you as individuals. I will answer the questions in this talk.
What are the trends? What could be the possible doomsday scenarios? Does AI actually replace people? What is the ideal world? Why do we need to emphasize about the black box approach and the white box approach in applying AI to human capital ? and What can you do as individuals to support the community and the ecosystem?
okay now let's talk about the trends. We will be superhumans in the next 5 to 10 years when AI empower us to do far more greater things than we can do the same things all by our own capabilities. We can achieve a lot of intellectual horsepower by working with them. One way to think about AI is that AI is a general purpose technology or GPT in short.
With reference to Bresnahan & Trajtenberg in 1995 in the Journal of Econometric, they wrote a paper called the engines of growth and they mentioned about the impact of the general purpose technology. These technology serve to create a leap or paradigm shift in the way human beings or society will operate and behave. Like any GPT, there will be this shift and transition. During this shift and transition, jobs will be lost and displaced. Task in the jobs will be automated. At the same time new jobs will be created. The ability to deploy the technology is often limited by how fast that the worker can be trained for it.
Let me show you a picture of what i mean. There are 2 circles in the picture. The grey circle represents the existing skill set of the worker while the red circle represents the new skill sets. There is interdependence of artificial intelligence adoption and the labor market conditions. In a market that doesn't have a new technology coming in , you, the worker, know that your existing skill sets and your required skill sets are pretty much the same, so the gap is very small or there is no gap at all. In other words, the grey circle superimposed itself on the red circle.
However when you have new technology coming into the market, you will notice that there is a need to require the new skill sets to operate the technology. The difference between the new skills set and the old skills set reveals the visible gap between them. And so you see that the overlap between the current skill set and the new skills set starts to widen. This is because we start to see that some workers become early adopters of the new technology and supply these new skills in the market. That’s when the technology evolves very fast to the point that the new skill sets almost does not overlap with the existing skill sets. As such, you will see that the red circle is moving to the right as more and more workers learn the new skills and more and more companies adopt the new technology. The intercept between the grey circle and the red circle become smaller and smaller over time. Over time, there will be an increasing need for reskilling.
And if the average worker does not become an early technology adopter, he will not be able to keep up in learning the technology. As these new technology keeps on adding more and more features, that worker will start to realize that the learning curve for these new technologies become steeper and steeper over time. That worker will find his or her employability falls over time.
Yet, why are we seeing a phenomenon that companies are reluctant to hire talents with emerging skills and even more reluctant to send talents for training? Why are we also seeing another phenomenon that individuals are skeptics about investing their time and resources to acquire new skills?
This is the reskilling paradox. Before I explain the reskilling paradox, let me explain what is reskilling.
Reskilling is often a time when the individuals needs to learn new skills and see that their existing skills get obsolete over time. For instance I would need to go and learn a set of new subjects. These subjects allow me to obtain the new knowledge. And then I need to practice the new knowledge and to develop the new skills. Thereafter, I will only then be able to offer the new skills to the market.
The first paradox of reskilling is that you can make more money using your current skills as compare to using your new skills when you expect to make more money on your new skills. This is because you are slower at applying your new skills to work relative to your current skills.
Now you learn the new skill. You are applying that new skills. What if you are applying that new skills at a rate that is far slower than the utilization rate of your existing skills to do the job? In other words, the reskilling utilization rate is slower than the old skill utilisation rate. let me give you an example.
For instance, let's say I go and get a job at a cafe. The old skill set is about taking orders and collecting cash just using a pen and paper. I can serve 50 customers using that old skill set.
Now suppose my boss in the cafe asks me to use an iPad or a tablet to take order and to administer the credit card payment using the tablet. I would need to learn to use the tablet and to learn to use the software in the tablet. Suppose I spend 40 hours to learn to use the tablet and the application in the tablet. That is the time cost of reskilling. And then I apply this new skills to serve the customer. Now let's also suppose that if I am not tech savvy like a millennial, i can only serve 25 customers using the iPad and the application in the iPad to collect payment and take orders.
So now when you think about using my old skill set to serve 50 customers and using my new skill set to serve 25 customers. That means I serve 25 less customers using the new skill sets relative to the old one. 25 customers can be a big financial loss to my boss and my boss would be reluctant to send me for more future training. This is because he realize that he is making a loss by sending me for training. This is the reskilling paradox. So this is what the market is concerned about. That is, the reskilling utilisation rate is slower than the old skill utilisation rate to the point that no companies wants to pay for training. If AI accelerates this paradox, then all the of the companies and business owners are concerned about the return of investment or ROI of investing in training.
And so what about the worker? Would the worker pay for his own training or invest in his own training? If he realizes that his new skill utilisation rate is slower than his old skill utilisation rate, then he would be smart to know that any future employer will notice that difference. That worker can predict that his future employee, upon noticing that difference can decide to either cut his pay or give him less working hours.
In this scenario, the worker will not invest in training, no matter how many grants are offered to him, how many encouragement is given to him and how frequent the media is publishing the benefits of training to him. This is an investment decision for the individual. He begins understand the second paradox of reskilling.
Now let's look at the second paradox of reskilling. You can make more money by focusing on learning less sets of skills as compare to focusing on learning many sets of skills. This is what we call the “Learn less, Make more” dilemma.
Here’s the side note: If the first paradox and second paradox occur at the same time, then the company will hire more experienced hire or will do more outsourcing to a marketplace instead of hiring a full time worker. When the company seek a gigster or a freelancer or a independent contractor, they will try to optimize this arrangement as much as possible. This is especially when there is no legal compliance to offer benefits, to offer insurance or to offer any forms of protections to the worker.
Ok let’s get back to talk about the “Learn less Make more” dilemma.
As you learn new skills, some of your old skills and your current skills get obsolete. According to research, our memory operates on the “Use-it-or-Lose-it” basis.
Let me give you an example. There is WhatsApp. There is slack chat. There is Wechat. There is hipchat and maybe another 10 more chat apps out there. If you use one of these applications very frequently, then you will be very good at using that specific app. That's no doubt about it and you will obtain productivity from that frequent use of that application. If you use many apps that does the same function, it is not possible to use them on the same intensity and frequency as you would spend that total amount of time on using 1 app. As such, you might not be good at using any of the apps at all. It is a honest reality check. That will be memory loss or memory decline in those apps because you don’t use them very frequently overtime. Some of these apps keep adding new features and it becomes very hard to learn so many apps over time. And these loss of skills or the decline of skills happen because of such memory loss or decline. This is known as deskilling.
When reskilling is slower than deskilling, then the companies would question the investment of sending this worker for training. In other words, if your ability to learn & apply the new technique is slower than your inability to repeat the same task on the same technology, then the companies would question the investment of sending this worker for training. This becomes difficult for the worker. He or she has family commitments like taking care for the family , taking care for the elders. The worker will take a long time to go for training and find a new job. He figures that he will prefer to profit from the short term by selling his current skills as a gigster, a freelancer. When the worker enters into such an arrangement, they will involuntarily be less prepared for financial planning or uncertainties like accidents, health hiccups. He begin to realise that he can no longer get jobs given his current skills are obsoleted.
This is why individuals like Elon Musk advocates Universal basic income. This is because when the technology revolution is so fast that when you put the context of reskilling and deskilling in dynamics, people have emotions of pain and uncertainty, and not everyone has sufficient income to trade their existing working for the time to study and to gain mastery. This is especially when they acknowledge their slow reskilling utilization rate of these new skills. So over time, there can be a situation in which the hatred for technology swells and we can experience a Luddites revolution.
The Luddites were a group of English textile workers and weavers in the 19th century who destroyed weaving machinery as a form of protest. The group was protesting the use of machinery in a "fraudulent and deceitful manner" to get around standard labour practices. Luddites feared that the time spent learning the skills of their craft would go to waste as machines would replace their role in the industry.
It is important to note that AI can exacerbates the reskilling-deskilling paradox and there the inability to manage the societal acceptance of AI and the societal management of the individual reskilling-deskilling transition can create social instability. It is an important truth that a lot of government , alot of political leaders and politicians have to recognise.
Let's go back to another trend. As automation accelerate the re-composition of jobs, jobs will become gigs. People who are good at a specific skill wouldn't want to stay in a full time job if companies are just gonna keep shoveling a lot of apps to their employees.
Suppose the employee is very good at using a specific software to develop a website. There are dozens of similar softwares doing the same task. The employee will question: “Why should i go and learn it? I am asking this question because I feel overwhelmed learning these softwares and using them have not been allowing me to gain any substantial productivity or substantial income? I might as well freelance. I might as well outsource my availability in the marketplace for that deployment of that specific software.”
This is where the danger lies. You can see the trends that individuals are beginning to say: “ Forget about taking a full time job! I just want to do a gig or freelance. I have less scope of work given i'm productive on using that specific software. I just need to keep finding more of the same assignments or gigs. If I am good at 3 to 4 skills sets to make some trades as a gigster or freelancer, then I will not prefer to take up 10 to 20 more skills sets in a corporate job. I might as well make more income in the short term as much as I can. This is because I feel that corporate jobs are also no longer secured.”
Eventually, you will notice that the only full time job holders will be the entrepreneurs, inventors, investors and complex gigs jugglers. Based on my personal experience, any of the above mentioned occupation holders will handle at least about 100 tasks a day. And the winners will be gigs capitalizers. They will become good at asking for gigs or using gigs. And as such, we will see today's job becoming a composition of tomorrow's gigs.
We will see that jobs get recompose every 6 months with new tasks or elements that cannot be automated.
When there is a new technology shift, companies will reorganize themselves. Companies will hire and fire people. They will typically fire people who are resistant to change. They will also hire new people. That dynamic of changes will cause morale disruption and unhappiness. And that's how modern and agile enterprise will work. You might be thinking why don’t companies cope with change by retraining the workers instead of restoring to hiring and firing? This is the answer.
Suppose they retrain the individual and that individual might not be able to handle the new task that is demanded by the market. As a result, that worker can drag the performance of the company and that company has to go into firing people again. And because of that, the individual will become smart to realize that: “I join any company and get fired every 2 years. This is because of the reorg needs to keep up or mistakes that the company leaders make. If that’s the case, then I might be better off becoming a gigster or a freelancer.” Therefore they enter into the marketplace, to target different businesses and to offer their specialized skills.
On the same cycle, the enterprise thinks: “Hey maybe we should go to the marketplace, and hire freelancers or gigster. Not because we want to pay them less, but we are able to get the required skills fast enough to meet the dynamic changes in the market!”
And so over time we see the following happens: the enterprise becomes leaner and leaner, the individual’s job becomes more and more transient and the transient marketplace becomes bigger and bigger. And this cycle will feed itself and perpetuates itself. Eventually, we might see that a job get recompose every 6 months with new elements that cannot be automated. We will realize that the stability of the job is no longer about staying in a job for 10 years. At the current time, the political leaders forecast that the individual will stay on the job for 2 years. Maybe because of this cycle, we foresee that in the next 5 to 10 years, a stable job only lasts 6 months to a year. The tenure of a rank and file corporate job will eventually approximate to the tenure of a gig or an assignment taken by a freelancer or a gigster. I term this the transient workforce cycle.[The only risk in life is not to take risks in your work]
And so you think about it. As the individual, the freelancer or gigster or a full time employee, you will be thinking like this: “How do I take a calculated risk in my next gig or job? So long I can take calculated risks, I am able to take care of myself. This is because every gig or job decision is an investment to decide my future lifetime income. Taking calculated risks means that I understand the concept of high risk high return. In other words, I, the individual, should expect a high return from a high risk job.
This is because the market is directly transferring part of the company’s risks to the workers. This is very different in the old days when, you follow a sequential life cycle of an adult. Let me explain abit more. About 10 to 20 years ago, you will follow a specific path of living. You take on formal education for 15 to 20 years. It means completing your primary school, your secondary school, your college and maybe your university. After completing your formal education, you take a good corporate job and you wiggle your way to be at the top of the corporate ladder. Out of that, you earn your income as you contribute to the society. And then you save enough to retire. However, this sequential life cycle of an adult has changed. So what has changed? How has the life cycle of an adult changes over time?
Recalling that as the enterprises get leaner and leaner, they get more nimble at outsourcing their jobs to independent contractors and to gigsters. The full time corporate talents will enter the transient workforce cycle and become gigsters overtime. These gigsters become very specialize at specific skills. They can be remote coders, content writers, data scientists and so on. Overtime, the marketplace arrangement between the gigsters and the enterprises reach a dynamic equilibrium.
As such, the new life cycle will no longer be sequential. You recall that your current skills set originates from your past formal education that gets obsoletes over time. Instead of making an income via the skill utilization of your current skills, you have to reskill to be able to keep up with the market as your income declines over time This is because you are selling your old skills as a gigster and your old skills declines in value over time. And because you are a gigster, your income is no longer staggered. In other words, you don't get paid when you are sick or when you cannot turn up for work. Your income is constantly dependent on your ability to utilize your skills in real time. Maybe in the past, you hold a corporate job and you only have to trade your time between work and non work commitments. In the past, the company will pay for your training and pay your salary when you go for training. Now, you will have to learn to optimize your time between completing delivery orders, reskilling and managing other non work commitments. The life path of an individual moves from being a sequential one to a concurrent one.
When that happens, then the notion of security is perhaps all about developing more capabilities and about translating that opportunities into income using technology. Income security will mean a passive streams of gigs and that retirement does not mean that you stop working. It means taking less gigs over time. You still have to work. You have to be so good at utilizing your skill sets that you can command a premium for that skill. By doing so,you don't have to worry about it. And so you think about it. You will realize that the individual is joining a new form of workforce called the transient workforce. This is also known as the uberization of jobs.
[Transient workforce + Uberization of jobs]
By 2020, companies will no longer be just hiring only full time employees but a mixture of contingent workers. These contingent workers can be temporary employees, part timers, casual workers, season workers, highly skilled remote workers, experts and independent contractors.
As such the modern agile companies will adapt to this transient workforce to enable themselves to be constantly innovative and relevant to the dynamic markets. As a consequent of that, these companies can therefore manage their human resources to be as cost effective , efficient and innovative as possible. This is what we call strategic flexible workforce management. And that will become an imperative for agile companies.
So you think about that. As companies adopt a transient workforce management approach and contingent workers have to manage their concurrent life cycles, the need for universal benefit income becomes visible. So why do we eventually need universal benefit income? This is because the human mind demands cognitive slack to learn and to be innovative or creative. You need cognitive slack whether you are in the corporate figuring what jobs to buy or whether you are the gigster trading your skills for income. You need the mental space and time to be able to think freely and easily. Only when you can think freely and easily, that's where the aha moments that ignite the creativity, the innovativeness in you. You also learn better over time.
If you don't have mental space to think about the things in your head, your mind will ramble like a monkey and wonder whether you can pay for your bills for tomorrow , whether you can support your family. When you have universal income, you will be thinking like this: “I don't have to worry about my income and I have the time and effort to take risks. I have more time to learn. I can take more calculated bets to gain social mobility.”
Let me give you an example.
Suppose you have 10 gigs, the truth is that you can please everybody sometime, you can please somebody every time but you cannot please everybody all the time. And because you are doing 10 jobs or more, the likelihood you will make mistakes increases over increasing number of jobs you take in a short period of time. This is even if you work with robots or AI applications, you will make mistakes. Out of 10 jobs, you will probably do well in 7 jobs and that's pretty good. If each job are equally paid, chances are you can still make a decent expected living. That is a form of stability for you in that case. Then you will have enough time to take that risks. Only when you can afford to take risks, you can make bigger bets. You can afford to take on skills that are more complex because you can have the potential upside to double or triple your income. You don't have to worry and therefore you can afford to learn something new.
Now let's look at another scenario, you don't do well in the 3 jobs out of the 10 jobs you do in 1 year and that 3 jobs derives 70% of your income. Now you can expect a sharp reduction in your income to the point that you have to constantly worry about your work, your family and your future. This constant worry can lead to depression or can lead to further cognitive decline or mistakes. That further exacerbates a vicious cycle that can send you to a downward spiral in your income and land you into an urban poverty status.
That's why we need universal income to enable the individual to transit through tough times especially when one starts to experience a sudden shock in one's economic status.
Having universal income at the back of the individual's mind, that same individual, who is in the same situation in the first scenario, can take bigger bets in taking up more complex skills or take bigger bets to take on more difficult assignments.
When each individual feels assured in the society, it enables inclusiveness and promote collaboration. There is a study done by Dan Ariely, one of the famous behavioral economist. When human beings think that life is not a zero sum game, they will collaborate. They will come together and they will work together. And that collaboration results in achieving an outcome that is greater than oneself. This is very important. The society will be better off with big wins.
This is the reason why we are seeing bigger and bigger amount of philanthropy dollars at work.
The rich people realize that even though their hard work is a contributing factor to their prosperity, there are many other contributing factors at play. These contributing factors can be as follows: The policemen is doing their jobs to make sure everything is safe, be it cyber crime or social crime. The teacher is trying to make sure that the next generation of adults have the relevant skill sets. Your friends and family are offering moral support.
And because of that, the rich man says: “ Thanks to the society. I manage to make it big and I'm so grateful that I want to share my prosperity with everyone including the marginal ones.”
And this is the trend that we are seeing right now. Rich people are giving back to the society because they feel grateful about it.
If the winners that doesn't give back to society in a capitalistic world, then the society will feel that it is playing a zero sum game. Nobody wants a zero sum game. In the same sense, when you have universal benefit income, and when everyone feels inclusive, it becomes natural that work becomes a calling or a vocation. Giving becomes a blessing.
When you feels assured and feels happy, your needs will rise along the Maslow hierarchy of needs. You will want to actualize your full potential as your basic needs of food, shelter and transport are met. You will start to ponder whether the work you do have meaning. Does it have meaning? Does it have impact? It's very often the desire to realize one’s meaning of life and the desire for one to achieve impact that drive the individual to take calculated risks. These desires also enable the individual to take the plunge to go through the roller coaster of life. Like having a vaccine injection, universal income can drive individual to be adventurous and to come out even stronger out of the risky decisions he or she makes. And therefore you feel that you want to contribute , you want to win and you want to give back. Working in that sense becomes fulfilling. It becomes your passion. When you make money by doing what you love to do, you don't feel that it is a job. You don't feel that it is a chore or a job.
That's why most people call household chores because most people don't like to do it. And when you love doing something, it is no longer called a chore but a leisure. You make leisure your job and that's the best thing in the world.
And therefore when you have money and you work long hours to satisfy your self actualization, you would not think about spending. By the time you have excess wealth, you will think about giving. And so when you give, you consider giving a blessing. And so work becomes a form of self dignification. It gives you a sense of dignity and a sense of pride instead of survival.
At the same time, we see that there are developed countries trying to experiment on setting up different forms of universal basic income. As the society provides basic income to everyone, the individual can receive this help. That help enables him or her to have some cognitive slack and a sense of security. As a result, everyone will become more collaborative.
Along the same line, when a country is in short of labor supply or talent supply, that country starts to calibrate pro-migrant schemes. Calibrating these schemes with universal income might enable the local to feel that: " It's ok. We need more people to come and contribute. These migrants can pay a fair additional tax to allow us to be able to come back to contribute to the society as we learn from them. We can also gain a fair share of their contribution to the new market. This is because now we are supporting each other. "
In recent times, we are seeing the rise of xenophobia or the rise of anti-migration. These mentioned backdrops can be a political cost to bring in migrants. As such, migrant schemes with universal income can potentially reduce the survival anxiety of the locals. This is echoed by Jeffrey Sachs a famous economist that acknowledge the impact of AI and Automation on job displacement and welcome universal income along with Elon Musk , the CEO of Tesla motor affirmation. Elon Musk mentioned that it will essentially be necessarily to introduce universal income as the era of Singularity is coming. The time will come when AI and Automation start to displace jobs faster than enabling job placements for those being displaced.
Now the other counter argument to giving universal income is that it might not be sustainable. So how sustainable can this be? The sustainability is often a calibrating balancing act of figuring out the optimal number of migrants with emerging skills, the optimal amount of universal income to the locals and the optimal amount of tax from companies. If companies take big bets by becoming leaner and generating more profits and more revenue per inhouse employee, then taxing on these revenue will be able to benefits the society. As a consequence of that, the society creates a sense of assurances for individuals to learn and to try new technological tools and businesses.
The other concerns about universal income is about balancing multi-generational issues. Suppose the government starts to give universal income, it is very likely to come from the budget instead of placing a bumpy tax claim on companies or foreigners. In order to think about the multi-generational issue, we need to think about the work belief and patterns of the future generations. The youngest generation in the current workforce is the millennial.[Behavior of future generations]
Millenial in the current workforce are more keen to work for experiences. They will be thinking along this line. If I take a gig or a part time job, I learn a new experience. I'm learning something new. I can give back and I can share especially using modern social media channels like instagram, wechat, snapchat and so on. Most often, these skills are not traditionally captured in the resumes. As such, these workers want to gain more visibly on the digital realm. The more visible the new age worker can showcase one's work, the higher the perceived personal brand one commands.
The notion of sharing and giving back without feeling a lot of effort in completing jobs make them feel cool and happy about it. In the digital world, they can be clueless given that they have so many choices.
They want to know what they are good at and what they can contribute. This fuels their desire for jobs with a lot of exposure. This can also means that they want to move to as many sectors as possible, to try as many job functions as much as possible and to go to as many countries as much as possible.
They want to deepen their transferable skills. In order to enable them to have more transferable skills, they will rather work in a company that enable them to share their portfolio or their working lives.
Companies like security companies or governments with classified information might not be able to allow the millenials to share their work informations, so they will be expecting to pay a higher salary premium relative to companies that enable that type of sharing. As such, restricted companies can expect to pay double or triple to hire a millenial. Companies can possibly resort to outsourcing to contractors. Independent contractors will probably take these jobs on a very selective basis.
The other interesting observation about the younger generation of workers is that they will work for a purpose. This means that they will get no pay or very little pay if they can tell people that "hey I do XYZ and it helps ten thousands people. It can helps 100,000 people." They think that their work has a greater impact and in order for that to happen, they would be thinking about what kind of work we should do to generate impact. Universal income can generate more volunteerism because more millenials are willing to volunteer given that they can generate impact without worrying about their basic needs. The above scenario is perhaps one of the merits for universal income because it encourages labor participation in social sectors which often experience a labor crunch. This is especially where these labor markets pose a high employment cost or are located in countries or cities with high cost of living. Some of these social sectors have jobs that are hard to do and hard to attract people to do. As such, universal income can allow some volunteers to take on these hard jobs given that these volunteers trade their working time to volunteer. When this happens, it enables the society to promote social cohesion and social collaboration.
What could be the worst case scenario if we are not seeing political leaders and companies moving forward to capitalize on this trend or to adapt to this trend? We might foresee a possible luddite revolution. Let’s look on the other extreme scenario of using AI. We have autonomous vehicles. It's great to have them in countries with a severe shortage of manpower. What if the dominant composition of the local workforce in that society are taxi drivers? This creates a huge unemployment in that society. What can these taxi drivers do to transit? The problem becomes acute especially when the taxi drivers will take a long time to find other jobs that require other types of intelligence. Why is that so? Their key skills originates from their spatial intelligence. Spatial intelligence is a form of human intelligence to manipulate moving objects in the 3D world. If autonomous car can replace tax drivers, there are also other autonomous vehicles that can replace other drivers. Can the displaced taxi drivers drive a van? Yes, driving the van is similar to driving the car because both use spatial intelligence. But they would not most likely get to drive the van because autonomous van is also coming. So if most of the transport companies are moving to adopt autonomous vehicles, then there would not be any jobs that require spatial intelligence anymore. If that is true, then the taxi drivers experience rapid deskilling. Then they need to add new skills to their human capital. Then the question is how long does it take for the taxi driver to learn a new skill that require other types of intelligence other than their spatial intelligence? Other skills like creative thinking ,digital marketing or programming, can take a long time for the displaced taxi drivers to learn because these new skills use other types of innate intelligence. It is like asking a swimmer to become a 100 metre sprinter when swimming is no longer an accepted sport in the olympic. And so when that happens, you might see a luddite revolution in that society. And that's what Jeffrey Sachs say that, “Yes. AI automate tasks but also create a lot of jobs. A lot of mini-tasks will be displaced.” And therefore there needs to be reskilling. If these taxi drivers have encountered the reskilling-deskilling paradox, then they not only cannot find a job, they will feel frustrated and possibly behave like the luddites. This is because of the sharp fall in their income and the sharp rise in their learning curve to pick another skills. These can propel a human being to behave irrationally or angrily. This is when universal income helps to smooth the transition for this group of people. If political leaders and companies ignore this trend, then they will face dangerous and volatile consequences. Now, let's look at another trend.[How can we find work for existing workers?]
There is a paradigm shift in which that the existing tools start to lose its relevance relative to emerging tools. Let me give you an example. Let's look at the job of a marketer. In the old days , a marketer has to decide on media buying from traditional publication sources. It can be buying ads from TV or buying ads from Radio station or buying billboards or buying flyers or buying the services to enable free sampling. Because of technology, companies are saying, “What is the return of investment for the marketing spent?” New technology tools like Google search or Facebook campaign start quantifying the marketing spent. And therefore digital marketing becomes a new scope of the existing marketers job. Here’s the question. Can existing marketers be able to transit themselves? These marketers are probably be thinking: “How difficult it is to transit? What is the time taken to transit? What is the economic cost for this transit? For example, will I have even less time to take care of my family? to take care of my kids? to be a good son or daughter to my elderly parents? What is the mental cost or emotional cost ? What is the cost for the government if the locals are not able to transit ?”
A Mckinsey research shows that the jobs that can be automated. The study describes the percentage of the tasks that can be automated in the jobs. The degree of automation is increasingly happening across all sorts of jobs over time.
If that's the case, what do we need to think about it? So how can we find work for the existing workers?
Here is an infographic about the hourly wage effects of automation on the jobs. As you can see the scatter plot is widely dispersed. So what does this mean? It means that automation happen to all the jobs. It also means that there is no linear or unique patterns for us to describe this relationship. Accounting jobs can be automated. Sales jobs can be automated. Food preparation in a restaurant can be automated. So any job that you can think of with an element of repetitiveness can be automated.
Let's look at another chart. If I have more and more roles in 1 job, then the probability of my job being fully automated becomes smaller and smaller. In other words, if I have a job consists of a data scientist and a marketers, then I would not be easily replaced by automation. Having said that, nobody wants to have a big job scope. No one wants to take on so many tasks. Taking a big job scope or taking a lot of tasks can very easily incur mental fatigue or burnout for you. Without having a sense of awareness, prolonged mental fatigue or burnout can lead to chronic health issues like mental illness or diabetes over time. These are the things that can cost the public lots of resources to fix it. Perhaps we need to start thinking about these questions. How can we enable our next generation to better prepare for automation across their jobs without worrying about long term health complications down the road? How can we better prepare the next generation to adapt and to beef up their mental resilience? And so you will feel like this: “Wow it seems so scary to think about it. AI can replace people or potentially increase workload expectations! AI can wreck my health.”
So how else that we human can do better ? Turns out that we still have some edge in the market for every human being. Let me share with you an example. A blu ray player at this time is almost 90% made by robots , ship by robotics arms and delivered by drones. It is currently sold at $47. At the same time, we have this handmade ceramic bowl that is 100% made by an artisan. It is currently sold at $750 in Ebay.
So you can see that the human authentic experience has a higher value than a 100% robotic solution. So as much as alot of jobs get automated, there are certain things that , we, human beings deliver or make in the form of services or goods exhibit a human touch and that human touch is highly valued by another human in the marketplace. And that is a great thing! At least we know that human generated value still carry premium over robotic generated solution.
Here's another example. Ok, you can pay $0.99 for an itune song or $10 a month on spotify for a song on Mozart played by a new artist. By the way, that song can also be composed with modern machine learning techniques. Yet you have to pay $807 for a live concert played by the same artist. The joke here is you probably pay $10 for a museum entry to watch the robot play the exact same song in real time. so clearly we see that there is a unique value generated between human and human connections. And complex communication becomes increasingly valuable. And this is one of the great things we need to take home.Job evolves because of AI
The other thing we need to think about AI on human capital is the AI effect on the evolution of existing jobs. How can we co-bot with the robots? How can we cooperate with the robots?
For example, we used to see the chef or hawker learning to operate the end to end kitchen activities chain by himself. There are a lot of tasks go into creating that special dish for the customer. As a chef, you have to source for the material, contact the supplier to make sure he or she comes on time, check the quality of the material, prepare the ingredients, chop them, slice them, sort them out, cook them, decor them, serve them and collect payment for them. It's almost 12 tasks just to serve a dish of chicken rice to the customer. Now with robotic chef, the number of tasks can be automated so the chef can do much less kitchen grind work and using that extra time to focus more on creating new recipes. This literally frees up the cognitive thinking of the chef. The chef becomes more creative and more interesting dishes will appear. New recipes and new ideas are formed faster. They are being tested faster until you see these chefs produce the Michelin standard. That's great because the rich people are willing to spend more on good food and the money gets circulated into the economy. And that's the beauty of AI.
Here's another example, the locksmith used to prepare the keys when someone locked himself or herself out of the house. That locksmith have to bring a bunch of keys to test. Now as more and more doors get digital locks, the role of the locksmith has evolved into a cyber security analyst that protects the security network of these digital locks. This is how a dying or obsoleting trade get evolves into a new trade.
Here's another example. An accountant that does bookkeeping and now he or she has to eventually evolve to become a forensic programmer. The task of automating the data entry and cognitive computing the 4 classical financial statements are now being done by AI. This is seen in software companies like Xero and Intuit. So the accountant has to spend the time generated by AI to focus on thinking about helping the customers to detect fraud or improve business performance.
The other impact from AI and Robotic is increasing the size of workforce for specific jobs. AI and Robotic can expand the talent pool for interesting jobs. Here's a cool example. From the old days till the current present, when you hire a police officer , you need to make sure that the candidate pass the physical fitness test. And then you deploy the new police officer. Now with robotics and AI, a kid that once lost his leg and arm can become a robot cop. His dream of becoming a police officer is now possible thanks to modern robotics and AI. This science fiction is becoming reality with new bio hack technology and equipments.
The other thing that I want to talk about is the nature of job. The nature of job will change overtime. Any job that is repetitive will be automated by AI. So what does that mean for you? If you want to survive, you have to take more and more different jobs. You have to take on more and more different tasks that is either hard to replicate or that is not repetitive in nature. These jobs typically require a lot of judgement, a lot of creativity, alot of innovation and alot of courage. When you try to do new things, your head or job is on the chopping board. If you don't muster your courage and try new staff with risky payoff, you are going to be mediocre. Your job will eventually be replaced by AI or Robotics. This is what we are seeing today. The demand for non routine cognitive jobs keeps increasing over time. At the same time, the routine jobs either get stagnated or are in decline. These are the key trends I want you to think about.Be a bit geek to increase job security
Now the next interesting key insights I want to share is about process thinking and the ability to apply that into automating your daily tasks. You want to be like a geek. What do i mean by that? Initially you might be doing a task. Let’s say you copy and paste in a certain way. You notice that you are copying and pasting on the same thing 10 times a day. It's so easy and it's so repetitive. If you can automate that, then you can clearly see the long run benefits. Yes, in the beginning, you will have to figure out the steps. You have to put them into writing and diagrams. Then you infer the algorithm and logic behind the steps. Then you figure a tool to automate that.
Yes, this takes a considerable amount of time in the short term because of the above initial activities that you have to make. You also feel the inertia to get started. You might have to use a lot of effort to automate that. However, as the number of repetitive tasks increases, you automate a lot of your tasks and that frees up alot of your time. You can use the free time to have cognitive free space to think, to be mindful and to learn and most importantly to innovate. And by doing so, you can try different gigs. In that way you are diversifying your careers risks and eventually transiting into a multi-stage life. And therefore be like a geek.
Now you have understood the above trends and their implications. how do you hack your career? how do you hack your job? Let me share with you a classical tip. If you find that your job is very routine, then the first thing you need to do is to network like crazy. That's network, network and network. Only via networking you get your name out there. You let people know about you and what you can do. Then you will receive different offers or gigs. That's when you can take as many gigs as possible. As soon as you can offer that same routine job to many different customers over time, then you are de-risking your job. In other words, you are diversifying your career risks. Your job that is routine might be automated if you stay in the same company overtime. But if you take the same job and offer to different companies, you might be productive at that 1 skill and make enough money to de-risk yourself over time.
What if your job is complex?
If your job is very complex, then you need to think about how can you learn faster , better and more effective. You will also need to think about how do you learn to focus. Because the truth is when you want to learn a complex issue, you need to take time to experiment and try a lot of things and that require a lot of focus. so the risk is the value of your job. So if you are unable to master that complex job because you are not able to perform that job and therefore you might lose that job, be it in the corporate environment or the outsourced environment. While the probability of losing this complex job is small, you have a high job value at risk. So what does that mean? For instance, you become the head of innovation. It is a complex job and few can take up this role because it requires design thinking, econometric thinking, ability to communicate and enable change management across the organization. In a typical job market, the demand for such candidates is more than the existing supply so the job fetches a high compensation package. Yet, once the candidate loses his or her job, he or she will lose a big drop in his or her income and that same person will take a longer time to find a similar job. This is why even professionals and high achievers should be advocating for universal income. This is because they will need to transit to another job and very often it's a long transit in terms of finding another job and adjusting to the sharp reduction in their income.
Now the next interesting concept that I want to share with you is about learning decisions. The decision to learn is actually as important as investing. The decision to learn is also the same as the decision to invest. If you go to a bank, a financial planner or relationship manager will ask you about your retirement plan.
It is far more important to think about the following: how do you plan your learning? how to learn? what to learn? from who to learn from? where to learn? how much time to learn? how much efforts it take to learn? what can you get from this learning? can you apply your learning to some practical use or futuristic use?
Because if you cannot learn to increase your potential market opportunities, then you will not have money to invest. This is because your potential market opportunities will predict your future lifetime income. If you know that your job is part of a cyclical trend, you can expect your fat paycheck to probably last for 18 months. If you know your fat income is that short, then you know that you need another 6 to 12 months to get a similar complex job. At the same time, you will not be able to invest regularly and have to draw on that saving to tide you to the next station.
As such, all the more you need to think about learning to hedge your career options.
If someone ask you about retirement, you should tell that guy to think about your learning strategies FIRST and then infer your investment options based on the dynamic state of your learning path and income implications from that path. You should ask that guy: what can you help me to learn? can you give me more network? so these are the things that I want you to think about it.
So what are the skills to learn? Because you can learn 101 skills like AI, Deep learning, digital marketing, cyber securities. You can also learn ecommerce. You can also learn video production. There are so many things to learn! But you only have 24 hours a day. It's not difference from what should i invest? there are stocks, bonds, safe deposits and many asset classes in the world. The recent mania is about bitcoins and ethereum on those ICOs. At the same time, there are many different kinds of fund managers with different investment objectives and investment strategies. If you can think like a fund manager on making a plan to learn and adjusting your learning portfolio, then you have a better chance of diversifying your reskilling risks and hedging against your deskilling risks.
There are 2 ways to think about this: “ What have i learn? Maybe I should be more sensitive to my own capabilities and deploying my time and resources to place my bet in taking this skill over the other skill. Which skill can I gain return of investment or ROI now?”
Because without quick wins, you will not have the mental motivation and the cognitive capabilities to keep learning and to keep learning very fast. That is very important for you to be aware of.
Recalling that I mention that we should process our learning decisions like investment decisions, then you might be wondering: “What are the principles of investing that can be apply to learning?”
Allow me to explain the concept of portfolio selection theory. This is developed by a Nobel economic prize winner , Harry Markowitz in his essay in 1952 on Portfolio selection theory. Suppose you have a limited budget. You can select different asset classes to reach your wealth planning objective, so you will eventually construct a portfolio that will give you an overall risk adjusted return. The classic principle states that the higher the risk you take in buying that asset, the higher the return you should demand from that asset. If we apply the same principles of investing to learning, then it means that you should take calculated risks in learning different courses and measuring the return of investment from your learning like an asset portfolio. In this case, you need to have a strong sense of self awareness on the skill utilization performance you get from learning the courses. By doing so, you improve your employability thereby improving your future earnings. This can be applied to everyone including myself.World of people powered AI Vision: Unlocking human capital using tech
Now let’s talk about a new world with AI that empowers human capital.
What does that world look like? That vision for us as a society is eventually to unlock human potential using technology. There are existing technology that can enable us to pursuit our happiness. This is similar to a Hollywood movie titled The Pursuit of Happyness played by Will Smith. We can eventually choose the type of work we want over time.
In 5 to 10 years time, companies will adopt new AI technology to evaluate the return of investment on human capital. The Moneyball movie will no longer be just a hollywood movie playing out in the sports arena but across all industries over time.
If we take a 10 to 20 years horizon, we might begin to see science fiction movies coming to reality. Like in the movie titled “Oblivion” played by Tom cruise in which human beings use advanced technology to clone themselves and transfer recent knowledge to their clones and sending them to different planets. Or it can be like the other movie “Ghost in the Shell”, in which human being can live forever when their physical body evolved using cybergenetics.
The next thing I want to talk about in today's context is the black box and white box approach.
What is the relevance of understanding these approaches in learning about human capital centric AI solutions? Why should we care?
In the white box model, you have very clear mechanics. You know what drives the outcome. It is typically econometric modeled. The white box model has very clear data driven processes and you can run through a series of validated hypothesis tests. You are able to interpret what drive the results. With the white box model, you can analyze and identify causality. This approach is very useful for building AI applications in the following HR sub-disciplines like managing transient workforce, running reward analytics, understanding the drivers of training effectiveness, optimizing onboarding practices, and enabling effective talent management.
On the other hand, we have the black box model which is an algorithmic approach that predicts outcome based on a bunch of data. It has an unknown data generation process so we cannot explain the mechanics of the pattern. It is very good to obtain a high level of accuracy. This approach is more suitable for building recommendation engines like recommending a good employee benefit, detecting unexplainable fraudulent behaviors and so on and so forth. In these applications, you don't really need to know what drive the behaviors or recommendation because these patterns are more often not explainable or sporadic in nature.
On the same note, the way to classify whether a machine learning technique is white box centric or black box centric is to look at its interpretability as well as its accuracy. The classical linear regression enables us to interpret the results so it's a white box approach in which you know the drivers of different outcomes but these regression models often times are hard to achieve high accurate prediction. On the other hand, the neural network models enable us to have very accurate results but we do not know what drive the results so it's a black box approach.Inductive Reasoning versus Deductive Reasoning
Now let's talk about the inductive reasoning and deductive reasoning. The way human beings solve a real problem in the world uses inductive reasoning and deductive reasoning.
Human beings typically observe some patterns and infer some generalization about those patterns.That is inductive reasoning. Once they get an estimate of the patterns that they observed, they tried to test the estimates against another fact that is true. When a human takes one fact and validated with another fact , they are using deductive reasoning.
AI that uses deep learning or black box model approach typically gives a prediction that is very accurate to the actual outcome given the variance in the given data and the actual data from the natural occurring processes is very small.
In other words, deep learning or black box model are good at deductive reasoning ONLY if the future outcome runs on a data generating process that is almost exact to the past.
Once the black box model gives an inaccurate answer, we have no clue on why that happen. We only know that there is a large variance between past data generating processes and future data generating processes.
So why am I sharing with you this? By understanding the limitation of the black box approach and knowing that AI applications can have a big impact on human capital investment, we must be careful about these AI applications that cannot explain the mechanics that are deployed to affect human capital investment decision making.What can you do?
Now let's wrap up this talk. We are aware that AI will change mankind. The implications are coming so we have to be prepared. Because we have different roles in our lives, we need to acknowledge that each unique role can response to the coming effects of AI and Automation on human capital investments.
In this talk, I share with you about the trends and hacks for the individual to be aware of. In the subsequent talks, I will share the different perspectives of other roles such as the government, the company and the investor.Let’s recap
Let’s recap about the tips that you have learned today. If you are taking a corporate job or trying out a gig, you need to understand that learning is like investing , consider the different income hacks for routine jobs and complex jobs. You need to take calculated risks in learning and taking on new jobs as singularity is coming.
Once again, thank you very much for your time.
This is Andrew Liew and we have come to the end of the talk.