It is not jobs that AI takes away
— AI takes away human jobs —
Recently, I hear that phrase often.
AI can now write text, create images, write programs, and even produce videos and music. It is not uncommon for it to finish in seconds tasks that used to take humans hours to complete.
That is why many people feel anxious that their jobs will disappear.
However, I feel a bit uncomfortable with that expression.
Is AI really taking away human jobs?
What is the difference between mechanization and AI?
Mechanization until now was very simple.
Returning the same answer to the same question over and over.
That is the area where machines excel.
Tightening the same screw.
Cutting the same part.
Performing the same calculation.
If the input is the same, the output is the same.
Therefore, through mechanization, variations in quality were reduced, and production volume increased dramatically.
On the other hand, AI is a little different.
AI answers different questions every time.
“Write an email about XX”
“Create a proposal for YY”
“Proofread the text for ZZ”
“Come up with ideas for AA”
Even though the input is different every time, AI returns an average answer that many people would find acceptable to that question.
In other words, what AI is good at is not automation.
It is averaging.
AI reduces variations in quality
In the manufacturing industry, there is a concept called “Six Sigma.”
It is an initiative to try to minimize variations in quality as much as possible.
Six Sigma (6σ) is a data-driven business improvement and quality management method that uses statistical techniques to suppress “variations” in business processes and minimize defects and failure rates (less than 3.4 per million opportunities), thereby simultaneously achieving customer satisfaction and improved management quality. “Sigma (σ)” is a statistical term (standard deviation) that indicates the degree of variation in data. It was developed at Motorola in the United States in the 1980s and spread throughout the world after General Electric (GE) introduced it company-wide and achieved great success.
When humans do work, there is inevitably unevenness in the output depending on their physical condition that day, their experience, and their strengths and weaknesses. However, AI can reduce that variation.
In other words, it is the image that while humans try hard to keep the output within a variation of about 3σ, AI can easily keep it within about 2σ.
Therefore, quality is stabilized, mistakes are reduced, and training costs are lowered.
From a company's perspective, it is ideal. That is why many companies issue a great order to “introduce AI.”
What AI takes away is not jobs
However, just because you can now produce consistent, uniform deliverables does not mean your job will change immediately. In fact, being able to produce uniform output is a good thing for you as well.
However, how the market views those deliverables will change.
Until now, there was a market that accepted deliverables even if they had some variation. However, once uniform deliverables become available at any time thanks to AI, the market standards themselves will change. There is no way that many people would intentionally choose to buy σ3 quality when there is an abundance of σ2 quality deliverables available.
This is not a story about humans losing to AI. It means
the market no longer requires that quality. That is why people feel like their "jobs have been taken by AI," but what is actually being lost is not the job itself, but
the market demand for those (σ2 to σ3) deliverables.
Who you should face
In today's world where AI is widespread, many people think, "I will make something a little better than AI." But that might just be competing in a quality range that the market no longer demands.
I do not think the thing you should really be thinking about is competing with AI.
What is the market looking for?
Who needs that value? It should be those questions.
It is not jobs that AI takes from people. It is demand.
The one we should be facing is not AI. It is
the change in demand.