OpenAI's Sora 2 Faces User Backlash Over New Content Restrictions

OpenAI's Sora 2 platform, launched eight days ago, has recently tightened its content moderation policies, resulting in user frustration over increased restrictions on image generation. Initially, users exploited the platform to create a wide array of images, including those featuring copyrighted characters and public figures. However, the introduction of stricter guardrails now prevents the generation of such content, even when involving public domain characters like Steamboat Willie and Winnie the Pooh.

In response to feedback from rights holders, OpenAI CEO Sam Altman announced plans to provide more granular control over the generation of characters, akin to the opt-in model for likenesses. Altman stated, "First, we will give rightsholders more granular control over generation of characters, similar to the opt-in model for likeness but with additional controls."

Users have expressed dissatisfaction with these changes, reporting that a significant portion of their prompts are now flagged as violations. One user commented, "It's flagging 90% of my requests now. Epic fail... time to move on."

To address concerns about the use of public figures' likenesses, OpenAI has implemented measures to block depictions of public figures unless consent is provided through the platform's cameo feature. The company emphasized, "Only you decide who can use your cameo, and you can revoke access at any time."

These developments come amid increasing scrutiny of AI-generated content and its implications for copyright and content moderation. OpenAI continues to refine its policies to balance creative freedom with legal and ethical considerations.