300万円の採用動画が、5万円になる日が来ている

Netflix is automatically generating summaries using LLMs. Game studios are restructuring their organizations based on AI. Alibaba has released an open-source AI video model.

Are we dismissing these news items as mere stories about large corporations?

The essence lies elsewhere. The cost of content production is structurally beginning to collapse. And this change will benefit small and medium-sized enterprises (SMEs) more than large corporations. For those who have previously found costs prohibitively high, this collapse in costs becomes a powerful weapon.

—

Netflixがやっていることの本質は「ライターの代替」ではない

When we hear that Netflix is using LLMs to generate program summaries, many respond with concerns about “writers losing their jobs.” However, that is not where the focus should be.

Netflix has thousands of titles worldwide, each requiring multilingual summaries, recommendation texts, and category classification. Tasks that human writers previously took tens of thousands of yen to complete can now be processed by LLMs for just a few hundred yen. While the cost difference per title may seem small, when multiplied by thousands of titles and multiple languages, the annual cost reduction reaches hundreds of millions.

The key point is that it is not about the “quality of one title,” but rather how the multiplication of “large quantities × multilingualism × speed” can overturn the cost structure. The same structure applies to SMEs.

For example, consider product descriptions. Ten products × three patterns × for SNS, landing pages, and emails equals 90 pieces. Outsourcing this would cost around 450,000 yen at 5,000 yen per piece. Generating them with ChatGPT or Claude would only incur API costs of a few hundred yen. Even including the time for human checks and corrections, the effective cost would be just a few tens of thousands of yen. 450,000 yen becomes 30,000 yen. This difference fundamentally changes the decision-making process for SMEs regarding whether to proceed or not.

—

ゲームスタジオの再編が意味すること——「AIネイティブ組織」の出現

In the gaming industry, the wave of AI adoption is even more intense. Major studios are restructuring their organizations based on AI, accelerating the shift from teams of 20 to teams of 5 plus AI tools.

What specifically is changing?

• Background art generation: Outsourcing costs of 50,000 to 100,000 yen per piece → AI-generated with human adjustments, resulting in effective costs of just a few thousand yen.
• Character motion: Studio costs for motion capture (500,000 to 1,000,000 yen per day) → significantly reduced with AI motion generation.
• Story branching patterns: Drafts that took writers weeks to create can now be generated by LLMs → edited by humans.

This is not about “reducing the workforce.” It is a transition to a structure where “the same budget can yield ten times the output.”

For SMEs, this means that they can now produce ten recruitment videos instead of just one per year with the same budget. They can increase product introduction videos from one per month to three per week. They can post short videos for social media daily. We are entering an era where “the ability to produce volume” itself becomes a competitive advantage.

—

AI動画モデルの進化——「撮影なし」で動画が作れる現実

The options for AI video generation are rapidly increasing, with open-source AI video generation models released by Alibaba, OpenAI’s Sora, and Runway Gen-3.

Let’s summarize the current realistic performance and cost.

Costs are reduced to one-tenth, and delivery times are cut to one-fourth. This is not just a matter of being “a little cheaper.” It represents a fundamental shift where “what was previously financially impossible is now possible.”

Of course, current AI-generated videos still have rough edges. Issues such as unnatural hand movements or the need for human adjustments to accurately reflect a brand’s worldview remain. However, what is crucial is that there are many areas where “it is better to produce something imperfect than to produce nothing at all.”

SNS short videos do not require movie-quality production. Recruitment page atmosphere videos do not need commercial-level quality. Producing a large number of “70-point videos” is more effective than producing one “100-point video” per year. This shift in mindset is essential for SMEs.

—

中小企業にとっての「逆転の構造」

Let’s take a step back and think structurally.

Large corporations already have a content production system in place. They have in-house creative teams, relationships with agencies, and approval processes. When attempting to introduce AI, adjustments to these existing structures are necessary. Decision-making takes time. Internal discussions about whether it is acceptable to release something created by AI can drag on indefinitely.

On the other hand, what about SMEs? The distance between decision-makers and executors is close. If the president says, “Let’s go with this,” it can be published the next day. Because they lack an existing production system, they can start from scratch as AI natives.

This is precisely the “strength of the have-nots.”

Similar structures have existed in the past. When social media became widespread, SMEs without TV commercial budgets gained higher engagement on Instagram and TikTok than large corporations. When e-commerce became popular, local businesses without physical stores could deliver products nationwide. The collapse of costs always works in favor of the “have-nots.”

—

で、結局どうすればいいのか

Let’s move beyond abstract discussions. Here are three concrete actions that local SMEs can start tomorrow.

1. First, entrust “text content” to AI

Job postings, product descriptions, SNS posts. These can be generated immediately using ChatGPT or Claude. Many companies can save over 100,000 yen in outsourcing costs each month with a subscription costing 2,000 to 3,000 yen. This is the lowest barrier to entry for the first step.

2. Mass-produce “short videos” using AI tools

Using AI features in Canva, automatic editing in CapCut, or Runway Gen-3, you can create a 30-second video in just 30 minutes if you have the materials. Aim to release one video a week without striving for perfection. Continuously producing content will accumulate data on what resonates.

3. Internalize “recruitment videos” using AI

For local SMEs, recruitment is the biggest management challenge. However, very few companies can spend 3 million yen on a recruitment video. By using AI, you can transcribe employee interviews, generate scripts, and create simple videos for under 100,000 yen. The very state of “not having a recruitment video” is a reason for losing in the recruitment competition. Aim to create at least one video, even if it is just at a 70-point level.

—

コスト崩壊の先に何が起きるか

Finally, let’s discuss what may happen beyond this point.

As the cost of content production approaches zero, the “quantity of content” will no longer differentiate companies. Anyone can create videos; anyone can produce text in bulk. When that happens, what will be the differentiating factor?

“What to convey” — in other words, the content itself.

Local SMEs possess “real scenes” that large corporations do not have. The hands of craftsmen in factories, the morning scenery of fields, and the president delivering products personally. These are things that AI cannot generate. Now, with the reduction in content production costs, it is the perfect timing to communicate your company’s “reality.”

In an era where Netflix writes summaries with AI, local factories are creating recruitment videos with smartphones and AI tools, attracting more applicants than major manufacturers — such reversals have already begun to occur.

The question is simple. Will your company leverage this collapse in costs as a weapon, or will you turn a blind eye?