The rapid pace of AI innovation often brings unforeseen challenges, and OpenAI’s groundbreaking Sora 2 AI video generation model is no exception. Launched on September 30, the model quickly ignited a significant debate around copyrighted content, leading to an immediate policy reversal. As Digital Tech Explorer, with insights from TechTalesLeo, continues to track the pulse of digital innovation, this swift shift highlights a critical turning point for AI ethics and creator rights.
In a blog post published on October 3, CEO Sam Altman directly addressed the wave of copyright concerns surrounding the new tool, announcing a pivotal policy update. OpenAI is now shifting its approach for using copyrighted material to an opt-in system. This means Sora will refuse prompts containing protected content unless the copyright holder explicitly grants OpenAI permission for its use, a move poised to reshape how AI interacts with existing intellectual property.
Japanese IP Infringement and Backlash
This policy change follows a significant international backlash, particularly from Japan. The internet was flooded with numerous Sora-generated videos featuring characters from popular Japanese games and anime, showcasing intellectual property from major entities such as Nintendo, Pokémon, One Piece, Demon Slayer, and Studio Ghibli. The controversy was perfectly encapsulated by one viral AI-generated clip showing a photorealistic Sam Altman surrounded by Pokémon, playfully stating, “I hope Nintendo doesn’t sue us.”
Tech enthusiasts and industry observers noted a stark contrast in how Sora treated different intellectual properties. While users reportedly struggled to create videos containing Disney or Marvel characters, Japanese IP did not appear to have the same automated protections. This discrepancy led to widespread criticism and calls for immediate action. Akihisa Shiozaki, a Japanese lawyer and House of Representatives member, highlighted the urgency, stating, “We need to take immediate action if we want to protect leading Japanese creators and the domestic content industry.” In his announcement, Altman specifically acknowledged the issue, noting, “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!”
Previous Copyright Stance and the Impetus for Change
Prior to this reversal, OpenAI’s policy for its Sora 2 model was based on an opt-out system for copyright holders. This approach effectively required creators to specifically inform OpenAI if they did not want their characters or work used by the AI. If a rights holder did not explicitly opt-out, their content was considered permissible for use in Sora prompts. This stood in direct contrast to the policy for real people, such as celebrities and public figures, which was opt-in by default, requiring their explicit consent. The significant legal exposure and the looming threat of copyright infringement lawsuits stemming from this policy difference is undoubtedly the primary reason OpenAI changed its position so quickly.
In his blog post, Altman further elaborated on the new direction, stating, “We will give rightsholders more granular control over generation of characters, similar to the opt-in model for likeness but with additional controls.” He framed the potential of the technology as a new form of “interactive fan fiction,” suggesting that many rights holders are “very excited” for this new type of engagement, which could add value to their IP. However, he acknowledged their paramount desire to control how their characters are used, including the crucial option to prohibit their use entirely. This shift reflects a growing recognition within the tech industry, a key focus for Digital Tech Explorer, that ethical considerations and creator rights must be foundational to advanced AI development.