OpenAI's video generation tool Sora is experiencing growing pains as the company scrambles to address copyright concerns and unexpected usage patterns. What started as an invite-only preview has quickly turned into a complex balancing act between user creativity and legal compliance.
The company recently announced significant changes to how Sora handles copyrighted characters and content. Users have been generating videos featuring everything from Disney characters to Nintendo's Mario, prompting swift action from rights holders. The community response suggests this wasn't entirely unexpected - many users saw the initial unrestricted access as too good to last.
Usage Patterns Exceed All Expectations
OpenAI admits they underestimated how much content users would create. With limits set at 100 videos per 24-hour period, users are burning through computational resources at an alarming rate. The high costs of video generation - estimated by some community members at several dollars per video - are creating serious financial pressure on the company.
The usage pattern has also surprised OpenAI in another way. Instead of creating viral content for public sharing, most users are generating personalized videos for small audiences or private use. This creates a mismatch between OpenAI's apparent expectation of a TikTok-style platform and the reality of users treating Sora as a personal video creation tool.
Current Sora Usage Limits:
- 100 video generations per rolling 24-hour period
- Invite-only access required
- 10-second video length typical
- 720p resolution standard
Revenue Sharing Model Emerges
To address both copyright concerns and financial sustainability, OpenAI plans to implement a revenue-sharing system with rights holders. This would allow companies like Disney or Nintendo to receive payment when users generate content featuring their characters. However, the technical and administrative complexity of such a system raises questions about its feasibility.
The community has noted a stark contrast in how different types of creators are treated. While major corporations may benefit from revenue-sharing agreements, individual creators and smaller content owners are likely to be left without similar protections or compensation opportunities.
Estimated Costs:
- Video generation compute cost: $4-5 USD per video (community estimate)
- Google Veo 3 comparison: $0.40 USD per second for users
- Sora videos typically 10 seconds each
Copyright Restrictions Spark User Backlash
Recent reports indicate that OpenAI has already begun implementing stricter content filters, preventing users from generating videos with popular copyrighted characters. The community reaction has been swift and negative, with many users expressing frustration that the service has lost much of its appeal without the ability to create content featuring recognizable characters.
These video generators are mostly useful for memes at this point, and adding copyright shackles make it a lot less useful for memeing.
The situation highlights a fundamental challenge for AI companies: balancing the creative freedom that makes their tools appealing with the legal requirements needed for sustainable business operations. As open-source alternatives emerge from companies less constrained by Western copyright laws, OpenAI may find itself at a competitive disadvantage.
The controversy reflects broader tensions in the AI industry about training data, fair use, and the rights of original creators. As Sora moves toward a public release, these issues will likely become even more prominent, potentially setting precedents for how AI-generated content is regulated and monetized across the industry.
Reference: Sora update #1