Safety first! Sora 2 users get new cameo control and content safety - Here's how

OpenAI has announced a major update to its AI video creation tool, Sora, enhancing user control over personal likenesses and improving content safety measures. Bill Peebles, the Head of Sora at OpenAI, shared these updates via X on Monday, following extensive feedback from users regarding content moderation and the handling of personal images. One of the most significant additions is the new control over how user-generated cameos can be utilized. Previously, users had limited influence over the appearance of their likeness in others' videos. Now, with the latest update, users can set specific restrictions, such as prohibiting their images from being associated with political content or banning certain words. Peebles emphasized that this enhancement is a move towards empowering users while still encouraging creative expression within the platform. Users can access these new settings through the menu: Edit Cameo > Cameo Preferences > Restrictions. Peebles also mentioned that OpenAI is committed to continuously improving moderation and customization tools for Sora. In addition to user controls, the update includes a revamped watermark system designed to be clearer and more visible, promoting transparency for AI-generated videos. Furthermore, OpenAI is implementing backend safety updates aimed at improving the detection of potential misuse and minimizing false negatives during moderation. Peebles acknowledged the challenges of over-moderation, noting that it can be frustrating for users. The update also addresses a long-standing issue: users can now delete their Sora accounts independently of their ChatGPT profiles. This change is expected to enhance user experience significantly. Peebles hinted at future developments for the platform, confirming that features allowing fictional characters to appear in cameos are being considered and will be detailed soon.