Conclusion First

The price of ‘voice’ is currently collapsing from both the top and the bottom.

The cost of generating AI voices is approaching nearly zero. On the other hand, the compensation costs for unauthorized use of someone else’s voice are skyrocketing. These two trends are happening simultaneously.

For small and medium-sized business owners, this is not just a matter of “let’s use it because it’s cheaper.” The structure of ‘saving 50,000 yen a month while taking on litigation risks worth millions’ is now right in front of us.

—

The Significance of ElevenLabs’ 22 Billion Dollar Valuation

AI voice synthesis startup ElevenLabs has doubled its valuation to approximately 22 billion dollars (about 3.3 trillion yen). This increase came just a few months after the last funding round.

This figure indicates that global investment money is flooding into the technology that “reproduces human voices.” Why? Because the cost of voice replication is dramatically decreasing, and the potential market is enormous.

Call centers, narration, customer service guidance, internal training videos, multilingual support—there are far more jobs that have relied on “human speech” than one might imagine. With ElevenLabs’ technology, it is possible to generate a clone voice that is nearly indistinguishable from the original using just a few seconds of audio samples. Moreover, it can support multiple languages. A voice recorded in Japanese can speak in English or Spanish as the “same person’s voice.”

Investors are looking at the potential of this technology to fundamentally change the massive cost structure of “voice labor.”

—

A World Where a 50,000 Yen Phone Operator Becomes 5,000 Yen

For small and medium-sized enterprises in rural areas, this change is even more immediate.

For instance, consider phone answering. If a small business uses a phone answering service, it typically costs around 10,000 to 50,000 yen per month. Hiring a part-time worker for phone duties would cost between 80,000 and 120,000 yen a month, while a full-time employee would cost over 200,000 yen.

Currently, AI voice agent phone answering services are beginning to emerge at around 5,000 to 10,000 yen per month. They offer 24/7 service, unlimited simultaneous calls, and they don’t take breaks or get moody. For FAQ-level inquiries, they can provide responses that are comparable to those of a human.

The same applies to video narration. Hiring a professional narrator for a 5-minute video can cost between 30,000 and 100,000 yen. With AI voices, the cost is nearly zero. For companies that produce internal training videos monthly, this results in a cost difference of several hundred thousand yen annually.

Costs are reduced to one-tenth. This is not just a “slight reduction”; it is a level of change that alters the criteria for deciding whether to proceed or not.

What was previously deemed “not cost-effective” has now transformed into “there’s no reason not to do it.” Multilingual product introduction videos, audio manuals, 24-hour phone support—what only large corporations could do is now possible for companies with just five employees.

—

The Other Collapse—The Price of ‘Voice Rights’ is Soaring

Up to this point, it could end with “let’s implement it right away.” However, another movement is occurring simultaneously.

In 2024, a noteworthy ruling was made by the Tokyo District Court. In a case where an actor’s voice was unauthorizedly replicated by AI, it was determined that the right of publicity extends to voices as well.

This is significant. Until now, Japanese law clearly prohibited the unauthorized commercial use of “faces” and “names,” but the legal status of “voice” was ambiguous. Now, the courts are beginning to move in the direction of protecting voices as well.

Similar movements are happening in the United States. Actress Scarlett Johansson recently protested against OpenAI for using a voice similar to hers, leading OpenAI to withdraw the relevant audio. In Tennessee, a state law known as the “ELVIS Act” has been enacted to prohibit the unauthorized replication of voices using AI.

In other words, while the ‘cost of creating voices’ is decreasing, the ‘consequences of unauthorized use of voices’ are increasing. These two vectors are moving in opposite directions.

—

Three Lines Small and Medium-Sized Enterprises Should Follow

So, what should small and medium-sized enterprises do? The worst decision is to think, “Let’s not do it because it’s scary.” Given that the cost structure is changing, companies that do not adopt this technology will lose their competitive edge.

The issue is “how to use it.” There are three lines that should be understood at this point.

1. Clarify ‘Whose Voice It Is’

Do not copy the voice of an actual person. This is the fundamental principle. Many AI voice services provide “original voices.” ElevenLabs also offers rights-cleared stock voices. Mimicking the voices of celebrities or narrators from competing companies is out of the question.

2. Read the Terms of Use (Really Read Them)

The terms of use for AI voice services outline “whether commercial use is allowed,” “the rights to generated audio,” and “prohibited uses.” If issues arise from using the service without reading these, the excuse of “I didn’t know” won’t hold up. Even a service costing 5,000 yen per month carries the same weight as a contract.

3. Do Not Hide That It Is ‘AI Voice’

If using AI voice for phone answering, clearly state at the beginning, “This call is being handled by AI.” This is not only to avoid legal risks but also a matter of trust with customers. The damage from being found out after hiding it is far greater.

—

The Real Challenge Lies Beyond the Lower Cost of Voice

Now we come to the main point.

The cost of phone answering with AI is decreasing. The cost of narration is decreasing. This alone will become accessible to any company. In other words, cost reduction will not be a differentiator.

What will differentiate is “what to do with the saved costs and time.”

Delegate phone answering to AI and use the freed-up manpower to increase customer visits. With narration costs at zero, produce weekly videos to enhance recruitment efforts. With multilingual support becoming cheaper, consider exhibiting at overseas trade shows.

The strength of small and medium-sized enterprises lies in their ability to act quickly from decision to execution. While large corporations are still processing approvals for AI voice adoption, small businesses can start using it as early as next week. This speed difference is the only structure that allows small and medium-sized enterprises to win against large corporations during technological transitions.

—

The Collapse of ‘Voice Prices’ is Just the Prologue

Next in line is video. AI video generation tools are already approaching practical levels. The era of creating “videos of people speaking” without the actual person is just around the corner.

When that happens, the sorting of “jobs that are meaningful for humans to do” and “jobs that can be entrusted to AI” will be necessary across all operations. Voice is at the forefront of this.

Reducing a 50,000 yen phone operator to 5,000 yen is just the entry point. Beyond that lies a survival strategy for small and medium-sized enterprises that focuses humans on tasks that only they can do.

I encourage you to try at least one AI voice service. Many offer free trials. By experimenting, you can determine within five minutes whether “this is usable” or “it’s still too early.”

What is needed for judgment is not commentary, but experience.