OpenAI is adding a series of important new control features to its still-developing Sora video generation app in response to criticism about the abuse of deepfake content.
According to Bill Peebles, the head of Sora, users will now have more power to define how their digital avatars are used in AI-generated content. Specific updates include:
Custom content filtering: Users can decide where their AI-generated avatars appear, such as blocking political content or prohibiting certain words from being used to describe or display their avatar.
Setting style guidelines: Users can set style guidelines for their digital avatars to ensure generated content aligns with their personal preferences and image requirements.
Support for Copyrighted Content and Revenue Sharing
In addition to enhanced user control features, Peebles also announced that Sora will soon officially support including copyrighted characters in videos.
This move aligns with recent statements from Sam Altman, CEO of OpenAI. Altman previously stated that copyright holders should have "more control" and revealed they will soon receive a portion of Sora's revenue, to build a more fair and compliant content ecosystem.
These updates mark Sora's steps toward enhancing safety, combating abuse, and establishing a clear copyright framework before its official launch.
