Will AI replace human labor?

For those who don’t know about AI, AI is short for Artificial Intelligence and its hottest related jobs are Machine Learning, Data Science, and Deep Learning.

(If you prefer short answer, scroll down to the bottom)

Hi guys, long time no see!

A little background from what I’m doing right now first.

Last month, I had the chance to learn “Applied Machine Learning in Economics” from VSSR (Vietnamese Summer School in Research) taught by Ph.D. from University of Birmingham (UK) and Harvard University (U.S). In terms of the technological side, this three-day course and one-day project obviously cannot cover everything in Machine Learning. What I appreciated the most, however, was my ML teacher’s insightful perspectives about Machine Learning in today’s world as “the one with data is the one with the power to control”.

Well, that’s actually correct! With data, companies will define people’s interests and predict their next goals for investments. With data, cooperations like Facebook, Youtube, or Lazada will control people’s consuming habits by continuously showing those “suggestions” on people’s posts. We machine learners and data scientists often joke about Son Tung M-TP, a very popular Vietnamese singer, as a data scientists not a singer due to the fact that he always knows his audience’s listening habits and produces music based on them.

Statistically proven, Data Science (or Machine Learning) is one of the hottest keys as well as skillsets in job postings for the past 10 years (and yet, it is still growing). What hurts me is to see some people with different passions, expertise give up their current jobs to devote themselves to Data Science for the sake of being recruited.

I mean, there is nothing wrong to learn about Tech, especially when it is so important in today’s development. However, it is not wise to risk your passion and expertise for something else for the sake of employment. Unless you have interests in Tech, your motives will be burn-out.

In fact, from my experience doing Machine Learning project with people from different fields. I learned that in order to acquire data sense, a Machine Learning team cannot ignore the help of experts who can give you what features needed to be eliminated or kept. For example, let’s just say you are solving a HR problem of selecting who will get a promotion using Machine Learning. Although you might gather all the possible features (years of experience, test score, age, level of education, etc.), only HR experts from that company will know what they need, which also means which features are useful to create a model.

We can generate a Machine Learning workflow to something similar to this, with requires the help of Experts, Data Scientists, and Machine Learning Engineer (Data Scientists can be replaced if Experts know how to utilize Machine Learning or if Machine Learning Engineer have data sense and the ability to analyze data):

  1. Define problems and goals. Experts, Data Scientists, and Machine Learning Engineer with finding the data together (preferred clean data, but that will be not the case in real life, so mostly requires data processing)

  2. Experts help eliminate some irrelevant features like ID, etc. (this depends)

  3. Data Scientists help visualize the data for better insights

  4. Experts help define useful features for training

  5. Finally, Machine Learning Engineers help train and test the model.

In other words, back to the main question “Will AI replace human works?” , I do not think so, not at this moment. Most Machine Learning works nowadays still requires support from experts due to the fact that Machine Learning/ Data Scientists are not always know-it-all people. Without experts, our models will be created based on pure assumptions.

But wait, if AI cannot replace human, what is happening? Why according to Fortune, “A.I. Expert Says Automation Could Replace 40% of Jobs in 15 Years” .

Theoretically, with the increasing computer’s capability according to Moore Law, most AI models nowadays can handle human works with greater automation’s capability in terms of dealing with large-scale data. In a more comprehensive way of explanation, computers/robots/AIs almost do not need to rest as much as a human but can still do a lot of automation or computation must faster than average human. That said, average workers, unless they learn to improve themselves, will mostly lose their jobs due to AI. This will inevitably happen as the world we are living in is not perfect; not everyone is a Ph.D. of their fields, concerning not everyone has chances to access education.

Short answer: Will AI replace human works? No, and at least at this moment when SAI (Super Artificial Intelligence) has not been born yet. AI models still need constant help from experts in their own fields to complete themselves, especially for some competitive tasks like generate music, artworks, self-driving car (in Vietnam laugh), etc.

However, AI will replace most average works that relate to automation. The only way we, as human, in this 4.0 industry can do is to improve our knowledge and become experts in our fields. Your experts’ works, directly or not, will definitely create some impacts for people from the Tech industry. In easier words, imagine AI like a baby child, if people from Tech are responsible to create a wealthy child’s body, the works of experts will be the soul and knowledge to enable that child to fully functional.

A message to non-tech people who are reading this post: Keep striving for knowledge and your expertise, your passion, things that can drive you forward in this competitive world. Don’t worry! If you are good enough, there will always rooms for your self-development.


Christopher Le