How to control your cameos: OpenAI’s Sora 2 update adds privacy, watermarks, and safety features

OpenAI has updated its AI video tool Sora, giving users more control over their cameos, improving watermark visibility, and enhancing safety. Announced by Bill Peebles on X, the update lets users set content restrictions, and delete accounts independently.

Govind Choudhary
Published7 Oct 2025, 08:14 AM IST
OpenAI is rolling out a significant update to its AI video tool Sora, introducing new controls for user “cameos,” improved watermark visibility, and enhanced safety mechanisms.
OpenAI is rolling out a significant update to its AI video tool Sora, introducing new controls for user “cameos,” improved watermark visibility, and enhanced safety mechanisms.(OpenAI)

OpenAI is rolling out a significant update to its AI video tool Sora, introducing new controls for user “cameos,” improved watermark visibility, and enhanced safety mechanisms. Bill Peebles, Head of Sora at OpenAI, announced the changes on X on Monday, after weeks of user feedback over how the tool handles personal likeness and content moderation.

Giving users more control over their digital likeness

One of the standout updates addresses a key user concern — how their cameos can be used by others. Sora’s cameo feature lets users generate AI videos featuring their likeness, but until now, they had limited say over how those likenesses appeared in others’ creations.

With the new update, users can impose specific restrictions — for instance, blocking political content or banning the use of certain words. Peebles described it as a step toward “keeping users in control” while allowing broad creative participation.

The controls can be found under Edit Cameo > Cameo Preferences > Restrictions, and Peebles noted that OpenAI plans to keep “hillclimbing” toward even more robust moderation and customisation tools.

Also Read | OpenAI’s Sora 2 sparks clone chaos: How to identify and download the real app
Also Read | Will OpenAI’s Sora rival TikTok or fuel more AI brain rot?
Also Read | From fan fiction to revenue sharing: Sam Altman outlines Sora’s next steps

Safety, watermarking, and moderation tweaks

Alongside user control features, OpenAI is refining Sora’s watermark system to make it “clearer and more visible,” reinforcing transparency around AI-generated content.

Peebles also confirmed backend safety updates designed to better detect potential misuse and reduce false negatives in moderation — while acknowledging that “overmoderation is super frustrating.” The company appears to be erring on the side of caution as it fine-tunes Sora’s safety systems in what Peebles called “a world still adjusting to this new technology.”

A practical fix is also included in this release: users can now delete their Sora accounts without having to delete their ChatGPT profiles, a long-standing pain point.

Alongside user control features, OpenAI is refining Sora’s watermark system to make it “clearer and more visible,” reinforcing transparency around AI-generated content.

Fictional cameos on the horizon?

The post also hinted at what’s next for the platform. When one X user asked if fictional characters could soon appear in cameos, Peebles confirmed the feature is “on the roadmap” with more details expected soon.

OpenaiArtifical Intelligence
Get Latest real-time updates

Catch all the Technology News and Updates on Live Mint. Download The Mint News App to get Daily Market Updates & Live Business News.

Business NewsTechnologyNewsHow to control your cameos: OpenAI’s Sora 2 update adds privacy, watermarks, and safety features
More