Robot Fear Strike at Hyundai Factory, Roblox Lets AI Create Games, Meta Sued for Using AI in Layoffs—’Losing Jobs to AI’ is Three Completely Different Issues

“Losing jobs to AI”

How many times have we heard this phrase? However, when we line up three recent incidents, we realize that while they are all described with the same words, the structure of the issues is completely different.

Thousands of workers went on strike at Hyundai’s factory. Roblox announced a feature that allows users to create games with just text input. Meta was sued for using AI in its layoff decisions.

All of these seem to be about “AI taking jobs.” However, the points that business leaders need to consider are entirely different. Mixing them up can lead to incorrect judgments. Let’s break them down one by one.

—

【Issue 1】Physical Replacement—Robot Strike at Hyundai Factory

In July 2025, a large-scale strike involving thousands of workers occurred at Hyundai’s automotive factory in South Korea. The cause was Hyundai’s plan to introduce the humanoid robot “Atlas,” developed by its subsidiary Boston Dynamics, into the production line. The fear among workers is that this robot, which stands about 183 cm tall and can lift over 45 kg, might physically replace their jobs.

This is the most straightforward structure of “tasks previously performed by human bodies being replaced by robotic bodies.”

The key point lies in “cost.” The introduction cost of industrial robots ranges from tens of millions to hundreds of millions of yen per unit. In large factories, this investment pays off. Hyundai’s Ulsan factory is one of the largest automotive factories in the world, with an annual production of about 1.5 million vehicles. Because of this scale, the amortization cost per robot can be lower than labor costs.

But what about small and medium-sized enterprises?

To be honest, the introduction of humanoid robots into small factories is still a distant prospect. The current introduction cost of Atlas has not been disclosed, but Boston Dynamics’ quadruped robot Spot costs about 11 million yen per unit. The humanoid will be even more expensive. For a factory with ten employees, it makes no sense to introduce such technology.

However, “still distant” does not mean “never coming.” Tesla’s Optimus has publicly stated a target price of $20,000 (about 3 million yen) for mass production. At 3 million yen, it is cheaper than one employee earning 4 million yen annually. If it operates 24 hours, productivity could triple. When this price point is realized, the landscape of small and medium-sized manufacturing will change dramatically.

What should be done now? It is not about buying robots. It is about taking stock of “what tasks can only be performed by human bodies in our workplace.” This will determine the accuracy of management decisions five years from now.

—

【Issue 2】Democratization of Skills—Roblox’s “Build” Feature

Roblox’s newly announced feature “Build” allows users to generate basic 3D games by simply entering text prompts. Until now, game development required specialized skills such as programming, 3D modeling, and level design. Now, it has become a matter of “just typing text.”

This is not a story about jobs being “taken away.” It is about the disappearance of the “barrier to entry” for jobs.

Let’s consider some specific numbers. Previously, outsourcing the creation of a simple mobile game would cost between 3 million and 5 million yen, with a development period of 3 to 6 months. With AI tools, a prototype can be created in a matter of hours, with costs approaching zero.

This structural change is not limited to the gaming industry.

• Website creation: Previously 500,000 yen. Now, less than 50,000 yen with AI.
• Video editing: Hiring a professional costs 100,000 yen per video. With AI tools, it’s less than 10,000 yen.
• Translation: Specialized translation costs 3,000 yen per page. With AI, it’s free in seconds.
• Logo design: Hiring a designer costs 100,000 yen. With AI, it’s a few hundred yen.

“What could only be done with specialized skills has become something anyone can do.” This is the essence of the second issue.

Is this a threat to small and medium-sized enterprises? On the contrary, it is nothing but an opportunity.

Large companies have specialized departments: design, marketing, IT. Small and medium-sized enterprises do not have these. Therefore, they have had no choice but to either outsource or give up. If AI breaks down the skill barriers, small and medium-sized enterprises will be the ones to benefit.

In a company with five employees, the sales representative can create proposal materials using AI, the accounting staff can update the website with AI, and the CEO can create video ads with AI. One person can perform three roles. This agility is something large companies cannot replicate.

The issue is not about being “taken away.” It is about the differences that arise from “not using” AI.

—

【Issue 3】Delegation of Judgment—Meta’s AI Layoff Lawsuit

The third issue is entirely different in nature. Meta was sued by former employees for using AI algorithms to select individuals for layoffs during a large-scale downsizing conducted in early 2025. The core of the lawsuit is that “AI analyzed personnel evaluation data to create a layoff list, but the logic behind it was opaque, resulting in an unfair number of employees with certain attributes being laid off.”

This is not about “AI taking jobs,” but rather about “when AI is entrusted with judgment, who takes responsibility?”

AI makes decisions based on statistical patterns. If there is bias in past personnel data, the AI’s judgments will also be biased. However, AI does not explain in a way that humans can understand “why that person was chosen.” As a result, the explanation “AI decided” does not hold up in court.

Is this an issue for small and medium-sized enterprises? Not at all.

Currently, there are services available that use AI for screening job applications, starting at a few tens of thousands of yen per month. There are also tools that analyze attendance data with AI to predict “employees at high risk of leaving.” These are convenient. However, when HR decisions are made based on “because AI said so,” the responsibility lies with the management.

Meta’s lawsuit made headlines due to its scale. However, similar structural issues can arise even in a company with 30 employees.

The lesson here is clear. AI can be used for “organizing information.” But “judgment” should be made by humans. The moment this line is blurred, risks emerge.

—

Don’t Mix the Three Issues

In summary, it looks like this:

The moment we lump everything under “losing jobs to AI,” our thinking stops.

Physical replacement is a cost issue. It may still be too early for small and medium-sized enterprises, but preparations can be made.

The democratization of skills can actually become a weapon for small and medium-sized enterprises. If something that used to cost 3 million yen can now be done for 50,000 yen, the gap with large companies narrows.

Delegation of judgment is a risk that needs immediate attention. There is a world of difference between letting AI “decide” and letting it “organize.”

What we should fear is not “losing jobs to AI,” but rather “lumping AI changes together and doing nothing.”

First, I encourage you to try AI tools in your own company. Whether it’s for proposal documents, image generation, or automating meeting minutes, using them will allow you to judge for yourself whether it is “taking away” or “becoming a weapon.”

You cannot act on abstract fears. Judgments can only arise from concrete experiences.