Teaching Your Job to AI for $12 an Hour

Hollywood screenwriters and producers are taking on gig work to teach AI their jobs. Compensation ranges from $12 to $200 per hour, approximately 1,800 yen to 30,000 yen.

They are selling off their long-honed expertise in screenplay structure, shooting schedule management, and character emotion design as training data for AI models.

The background is simple. With the rise of streaming, the number of productions has increased, but budgets per project have decreased. Even after the strike, work has not fully returned. Some individuals have found themselves compelled to choose “teaching AI” as a means of making a living.

However, there is a structural contradiction here.

The more you teach, the less work you have.

If AI learns how to write scripts, the next order will go to AI. They are digging their own professional graves at $12 an hour—something they are aware of. They are doing it, knowing full well, to pay this month’s rent.

This issue is not confined to Hollywood.

80% of Developers Feel “Addicted to AI but Unproductive”

There is another crucial piece of data to consider.

A recent survey revealed that about 80% of software developers feel a high dependency on AI coding tools. At the same time, many of them feel that “productivity has not actually improved.”

In other words, they can’t help but use it, but it doesn’t significantly improve their work.

This is a typical dependency structure. They review and modify code generated by AI, and in the back-and-forth, time slips away. There are many cases where writing it themselves would have been faster.

Moreover, there is a growing “fatigue” towards content generated by AI. On social media and blogs, the moment people recognize “this was written by AI,” they tend to ignore it. While the quantity has exploded, the amount that gets read has decreased.

Let’s summarize what’s happening here:

• The teachers: Transfer their skills to AI, making themselves unnecessary.
• The users: Depend on AI but become exhausted without achieving results.

Neither side is having a successful relationship with AI.

The Paradox of “The Teachers Disappearing First”

Here lies a paradox.

Commonly, it has been thought that those who will lose their jobs to AI are “people doing simple tasks.” However, the ones most affected first are actually experts who can teach AI their jobs.

Why is that? There are three reasons.

1. Specialized knowledge that can be “verbalized” is more easily absorbed by AI.

Three-act structure of scripts, theories of shooting schedules, best practices in coding—these can be taught precisely because they are systematized. And because they are systematized, they are easier for AI to learn.

2. The act of “teaching” itself creates its own replacement.

This is a different dimension from manual creation in the past. Manuals were read by humans and executed by humans. However, AI training data is read by AI and executed by AI. The moment you teach, your copy is born.

3. The cost structure changes dramatically.

If you hire a Hollywood screenwriter for a script, it costs tens of thousands of dollars. If you have AI generate it, the cost is almost zero. Knowledge taught at $200 per hour is implemented by AI at $0 per hour. This asymmetry structurally eliminates the “teachers.”

This is not just about a job worth 3 million yen becoming worth 50,000 yen. It signifies that we are standing at the threshold of a world where a 3 million yen job becomes worth zero.

So, What Should Small Businesses Do?

Now, let’s get to the main point.

The Hollywood story may seem distant. However, the structure is the same. In local small businesses, formalizing the tacit knowledge of veteran employees and systematizing operations has the same structure as “teaching AI their jobs.”

So, what can be done to ensure that the “teachers” do not disappear?

1. Become the Designer of the “End Result” of Teaching

The important thing is not what to teach AI, but how to design the use of the results of what you taught AI.

For example, automating sales pitches by teaching AI. Anyone can do this. However, the design of decisions such as “which customer to approach, at what timing, and with which pitch” can only be done by someone who knows the field.

Become someone who designs systems using AI, rather than someone who teaches AI their job. This is the first condition.

2. Have Areas That Cannot Be “Verbalized”

Skills that can be systematized are absorbed by AI. Conversely, those that are difficult to systematize will remain.

Local small businesses possess contextual judgment like “the president of that client has this kind of personality, so I should approach them this way” or “this region will have this kind of demand during this period”—these context-dependent judgments are difficult to dataize.

Large corporations run generic AI on vast amounts of data. Small businesses can compete in local contexts and human relationships, areas where AI struggles the most. This is not a weakness but a structural strength.

3. Use AI Cost Reductions to Be Proactive

When something becomes cheaper due to AI, many companies stop at “cost reduction.” However, the real differentiation comes from what to reinvest the saved costs and time into.

For example, if automating the creation of estimates with AI saves 20 hours a month, will those 20 hours be spent on “visiting existing customers” or on “experimenting with new businesses”? This is where managerial judgment lies.

Redirect the costs reduced by AI to things that only humans can do. Whether this judgment can be made will be a dividing line for small businesses.

From “Teachers” to “Questioners”

Finally, I would like to present one perspective.

Hollywood creators are teaching AI “answers.” How to write scripts, how to organize schedules—all answers that say, “this is how you do it.”

However, what AI truly struggles with is not “answers” but “questions.”

“Who is this movie made for?” “Is this product really needed?” “What will our company rely on to make a living in ten years?”—the ability to pose such questions is something AI lacks.

Those who teach answers can be replaced by AI. Those who pose questions cannot be replaced.

This is good news for small business owners. Because those who ask themselves daily, “Is this okay as it is?” are in the position that is least likely to be replaced by AI.

Will you sell answers for $12 an hour? Or will you position yourself as someone who poses questions?

As the evolution of AI accelerates, this choice is now before us.