OpenAI Sparks Major Controversy in Japan as Sora Copies Nintendo and Anime

Image: Legendary, Warner Bros./The Pokémon Company

OpenAI’s new generative video tool, Sora, has been public for only a week but already ignited a wave of controversy. Early users have pushed the app toward recognizable aesthetics and characters from major anime and gaming franchises, producing a flood of derivative clips and prompting widespread debate. PC Gamer documents several of those viral outputs.

Judging by the initial rollout, OpenAI did not include strong safeguards to prevent copyrighted material from being used as stylistic or narrative source material. That shortfall has already prompted CEO Sam Altman to outline forthcoming guardrails and controls for Sora.

Like ChatGPT but for moving images, Sora translates text prompts into video. Its capabilities range from convincingly recreating the look of established visual styles to producing realistic deepfakes of public figures. One widely circulated clip shows a photorealistic Pikachu interacting with a simulated Sam Altman; another appears to recreate an entire mission from Cyberpunk 2077, matching level geometry, vehicle designs, and voice characteristics with unnerving specificity.

Those outputs have raised alarm about misinformation risks and possible security implications when synthetic footage portrays real people. Observers are also asking how Sora attains such fidelity — specifically what datasets or assets OpenAI used to train or tune the model. According to the Wall Street Journal, OpenAI is now inviting copyright owners to opt out of having their properties generated by Sora. Some large studios appear to have enforced stricter blocks inside the app, especially for properties owned by companies like Marvel and Disney. The Wall Street Journal covers the developing policy.

By contrast, a flood of Japanese cultural content has been generated with apparent ease: Sora can mimic the visual language of anime such as Attack on Titan, Dragon Ball Z, and One Piece, often with strikingly faithful results. That has drawn public criticism and even political attention; clips and posts have spurred discussions among creators and legislators in Japan. One high-profile thread on X highlights the backlash and the concern that such technology could erode cultural and commercial control over beloved works. See the discussion.

In an update on his blog, Altman acknowledged that the public’s use of Sora has exposed issues the company hadn’t fully anticipated and said OpenAI plans to give rights holders more precise controls than a simple opt-in or opt-out. Altman’s post frames the changes as iterative: OpenAI intends to offer mechanisms for owners to specify permitted uses — including the option to prohibit usage entirely — and to apply those standards consistently across properties.

Altman says OpenAI will provide more granular options so creators and companies can decide how their characters and worlds may be used, and that the system will be refined over time to reduce unintended generations.

Nintendo, a frequent enforcer of its intellectual property rights, has been a recurring example in the conversation. AI-generated clips joking about the company’s potential reaction — including a viral short of Altman quipping “I hope Nintendo doesn’t sue us” — amplified the debate. Nintendo has denied reports that it coordinated with the Japanese government to block generative AI, while reiterating that it will take action against unauthorized use of its IP. Video Games Chronicle reported on the denial, and Nintendo’s own post restated its commitment to protecting its creations. Nintendo statement.

Source: Polygon