The arrival of Sora 2 not only represents a technical leap in the creation of videos with AI, it inaugurates a new territory of presence.

Among the advances in physics, sound, and consistency, one feature in particular has caught attention: the Cameo, which allows for the insertion of real people's faces and voices into artificially generated scenes.

What is Cameo and how does it work

The Cameo is a functionality available exclusively in the official Sora 2 app, currently restricted to iOS devices.

It uses a short reference recording of about 10 to 20 seconds, in which the user speaks a few simple phrases and makes small facial movements. From this material, the model is capable of:

• Mapping facial proportions and muscle structure

• Extracting the tone, rhythm, and intonation of the voice

• Capturing micro-expressions and natural speech patterns

This data is then applied to any video generated with Sora 2, replicating the face and voice with high fidelity.

The creator can choose in which scene, moment, and context they want to insert their presence, whether acting directly in a take or narrating a complete video.

💡 Important: Cameo is not just a facial filter. It reconstructs the face and audio within the 3D environment simulated by the model, maintaining coherence with light, sound, and movement.

Presence and representation

So far, AI video models have mostly worked with artificial characters. Cameo changes this logic by bringing the real body into the generation, incorporating human nuances, such as expression, rhythm, pause, and breathing of real people.

With this, Sora 2 moves from merely "simulating" people to representing them. This change shifts the focus from technique to authorship: who is present in the scene, and with what intention?

This is not about replacing human performance but expanding forms of participation. The face becomes yet another narrative tool, and the creator gains a new dimension of expression.

How to use Cameo in Sora 2 in 5 steps

1 - Record your reference material

• Use good lighting and a frontal frame.

• Speak naturally, with small variations in expression and head movement.

• The audio of the speech is essential; it helps the model reproduce rhythm and breathing.

2 - Upload the video in the Sora 2 app

• In the main panel, select the “Cameo” tab.

• Upload your recording and wait for processing (generally 2 to 3 minutes).

3 - Choose how to apply it

• In any prompt, add the command Cameo: followed by the name of your reference.

• Example: “Scene of a black and white perfume commercial, close-up on the character's face (Cameo: João), soft side light, introspective tone.”

4 - Control the behavior

• You can set the duration of the appearance, the intensity of the voice, and even predominant expressions (“smiling”, “neutral”, “worried”).

• The model automatically synchronizes the lips and tone of speech to the generated sound.

5 - Review and adjust

• After rendering, you can use the Remix tool to change light, angle, timing, or emotion without needing to redo the Cameo.

Current features and limitations

• Fidelity: the model maintains about 85–90% similarity to the real face, adjusting slightly to adapt to the lighting and lens of the video.

• Resolution: available at up to 1280x720, with optimized rendering for medium shots and close-ups.

• Duration: Cameo works best in short scenes (4 to 8 seconds). In long takes, synchronization tends to degrade.

• Audio: the system replicates intonation and rhythm, but does not exactly reproduce the timbre. It creates an “AI version” of the original voice.

• Privacy: the face and voice used are saved locally on the app and are not automatically sent to OpenAI, according to the official statement.

New questions about digital identity

The ability to replicate real identities also brings challenges. When a personal image can be recreated with accuracy, important ethical questions arise:

• How to ensure consent and control of use?

• Who is the author of a performance created by AI?

• How do these representations affect the public perception of those depicted?

These are questions that still do not have definitive answers, but need to accompany the adoption of the technology.

Cameo expands the field of creation and, at the same time, demands responsibility and transparency in how we use faces and voices within the digital environment.

Creative potential

Used consciously, Cameo offers new possibilities for audiovisual production:

• Independent creators can appear in their own videos without relying on filming.

• Brands and artists can create personalized narratives with total aesthetic control.

• Educators, performers, and communicators can explore presence and discourse in new ways.

The tool does not eliminate creative work; it just changes its nature. The challenge becomes directing oneself within a generated space, maintaining coherence and intention.

Conclusion

Cameo inaugurates a more complex stage of AI creation, wherein the human is not only the author but also becomes the material of creation. Sora 2 reinforces a central idea: it is not technology that defines authorship, but how we choose to be within it. 

Digital identity, once an extension of who we are, now becomes an active part of the visual composition. And in this new scenario, directing with awareness may be the main act of creation.