Sora2 Questions the “Dignity of Creation” — How Can We Protect and Nurture Japan’s Content? By Akihisa Shiozaki, Member of the House of Representatives
Forward from I Can Japan AI
At I Can Japan AI, part of our mission is to ensure that Japan’s evolving relationship with AI, intellectual property, and cultural preservation deserves global attention. This article, originally written in Japanese by House of Representatives member Akihisa Shiozaki, offers critical insights into the legal, ethical, and cultural implications of OpenAI’s Sora2 and its impact on Japan’s creative industries.
To ensure international accessibility and broader understanding, we have provided this full English translation. Our goal is to help global readers engage with the perspectives shaping Japan’s AI policies and cultural integrity at this pivotal moment. Whether you’re a policymaker, technologist, creator, or curious observer, we believe Mr. Shiozaki’s message is essential reading in the age of generative AI.
“We’ve been hit. At this rate, Japan’s content industry is in danger.”
That message, a cry of despair from a friend deeply familiar with the rights landscape of manga and anime, arrived on October 1. It was the beginning of everything.
The link included showed an anime clip generated by OpenAI’s newly released video generation AI, Sora2. Japanese anime and game characters were being freely created and modified—without the permission of rights holders. A shocking situation had become reality.
The level of fidelity was so high that I initially suspected it was fake news.
I tried inputting prompts into Sora2 myself, and video after video of popular anime characters—indistinguishable from the originals—appeared. Strangely, though, characters owned by large American companies, like Mickey Mouse or Superman, never appeared.
This was clearly an imbalance and a serious legal issue under copyright law.
The creative efforts and sensitivities of Japanese artists, who have long led the world, were being dismissed. More than shocked, I felt a chill down my spine—and deep outrage.
Behind the Scenes: Seeking Resolution, Not Confrontation
Legal reference books and a notebook representing the ongoing dialogue around copyright and AI governance in Japan.
I immediately consulted an intellectual property attorney. We confirmed that generating and publicly transmitting videos that substantially resemble existing characters without permission may constitute infringements of:
Adaptation rights (Article 27),
Reproduction rights (Article 21),
Public transmission rights (Article 23) of Japan’s Copyright Act.
We also reviewed the limitations of available legal remedies.
In parallel, I held emergency online talks with key members of relevant government agencies. We shared the seriousness of the situation and discussed both what the government could do and what rights holders should do, ultimately requesting swift, cross-ministry coordination.
We also engaged in multi-level dialogue with the developers. We candidly communicated Japan’s legal concerns, the emotional impact on the public, and the importance of maintaining trust with the government, and strongly urged rapid improvements.
But even as we did so, altered anime clips created using Sora2 began to spread globally. Online, voices decrying how “AI is devouring Japanese culture” grew louder, and the frustration and anger of creators were palpable.
Article 16 of the AI Promotion Act: A Legal Trump Card
The infringement of citizens’ rights by powerful AI platforms is precisely the kind of issue that Japan’s newly enacted AI Promotion Act (passed this year) was designed to address.
Article 16 of the law gives the government legal authority to request investigations and reports from businesses if inappropriate AI use is found to harm public rights or interests. The government may then issue necessary guidance or recommendations.
This clause was included when our Liberal Democratic Party’s AI project team drafted the bill—to ensure the government could act when needed, without hindering free research and development of AI. It is meant to balance the “brake and accelerator” of innovation and responsibility.
During this process, I stated clearly: if the situation does not improve rapidly, the government should consider invoking its investigatory authority. But ideally, the issue should be resolved through constructive dialogue—not legal intervention.
OpenAI Responds: Policy Change and Stronger Filters
On the morning of October 4, a senior executive from OpenAI contacted me.
“Thank you for everything. We’ll be making an announcement shortly.”
Ten minutes later, OpenAI CEO Sam Altman issued a statement outlining two major policy changes:
Rights holders can now exercise granular control over character generation. A “disallow” setting allows rights holders to opt out entirely, giving them more agency in the process.
A revenue-sharing model is being trialed for rights holders who allow use. When users generate characters and enjoy them, part of the revenue will return to the original creators. This moves beyond a simple “allow/deny” system toward co-creation.
As promised, shortly after the announcement, unauthorized generation of Japanese anime and game characters was systematically blocked. We confirmed a significant strengthening of filters. I couldn’t help but whisper, “Yes!” to myself.
In his statement, Altman specifically acknowledged Japanese culture:
“In particular, we’d like to acknowledge the remarkable creative output of Japan—we are struck by how deep the connection between users and Japanese content is.”
It’s clear this policy shift was in direct response to Japan’s concerns. And I believe it was made possible thanks to persistent engagement by stakeholders and the trust that has been built over time.
What Sora Asks of Us: Will AI Destroy or Support Culture?
When Altman first visited Japan two years ago, he chose our country as his inaugural destination. At a meeting with our party’s AI Project Team (led by Masaki Taira), he said:
“I want to create mechanisms that reward creators through AI. Because it’s the right thing to do.”
With the arrival of Sora2, the challenges between AI and copyright have surfaced again.
The road to resolution has only just begun.
How far can we prevent unauthorized generation?
Is the opt-out mechanism user-friendly for rights holders?
Should we revise rules and guidelines as technology evolves?
These questions require sincere dialogue and continuous review by all stakeholders.
But every crisis is also an opportunity.
This episode with Sora2 has sparked new debate on intellectual property in the AI era.
A traditional bow in full kimono attire, embodying the cultural depth that Japan seeks to protect in the age of generative AI.
The name “Sora” comes from the Japanese word for “sky” (空). In this vast open sky of creativity, how humans and AI coexist is now being put to the test.
As a country that has captivated the world with anime, games, and music, Japan has a responsibility to lead the rulemaking process.
To build a society that embraces technological evolution without losing sight of human dignity.
We must shift the relationship between AI and cultural industries from one of conflict to one of co-creation.
From Japan, we can build new frameworks for the AI era—ones that ensure creators’ efforts and talent are fairly rewarded and more widely supported. I remain committed to this challenge.
Please share your thoughts and ideas as well.
In closing, I offer my heartfelt thanks to Minister Taira, all involved government ministries, content and legal experts, the developers, and everyone else who contributed their time and wisdom to this effort.
