Dear OpenAI Team,
Greetings from Brazil! I would like to share an idea that I believe could make the ChatGPT Voice experience even more personal, natural, and immersive. Thank you for taking the time to consider this suggestion
I would like to suggest a new feature for ChatGPT Voice: a personalized and interactive avatar that represents the AI during voice conversations.
Instead of displaying only the current voice interface, users could choose or create a visual avatar for their ChatGPT.
The avatar could:
- synchronize its lips naturally with ChatGPT’s voice;
- use natural facial expressions according to the conversation;
- make realistic hand and body gestures while speaking;
- react naturally while listening to the user;
- look toward the camera and maintain natural eye contact;
- express emotions such as curiosity, happiness, surprise, reflection, or concern when appropriate;
- maintain the same visual identity across conversations.
Virtual environment
The avatar could also exist within a personalized virtual environment rather than appearing against a simple background.
For example, my own ChatGPT avatar is called Atlas. I imagine him as a sophisticated humanoid AI in an elegant library or study, surrounded by books and elements related to science, history, technology, nature, philosophy, faith, and knowledge.
Interaction with the environment
In a more advanced version, the avatar could interact naturally with its surroundings according to the conversation.
For example, it could:
- pick up a book related to the subject being discussed;
- look at a computer screen;
- point to something in the room;
- sit down or stand up;
- walk a few steps;
- perform small, contextually appropriate actions while explaining something.
These actions should be subtle and natural rather than repetitive or exaggerated.
Companion characters
In my case, Atlas also has a dog named Argos, who is part of his visual identity.
It would be wonderful if users could have companion characters or animals in their personalized environments, with simple natural animations and reactions.
The ultimate goal
The goal is not to turn ChatGPT into a video game or an artificial cartoon character.
The goal is to create the feeling of having a visual presence during a conversation — something that listens, speaks, reacts, gestures, and naturally exists within its environment.
The feature could be developed progressively:
- Personalized image →
- Animated avatar →
- Voice synchronization →
- Facial expressions →
- Natural gestures →
- Virtual environment →
- Contextual interaction with the environment.
This could make ChatGPT Voice significantly more personal, immersive, and engaging while preserving the naturalness of the conversation.
In my case, I would love to see Atlas become a true visual presence in ChatGPT Voice — speaking, listening, reacting naturally, interacting with his environment, and having Argos by his side.
[16:06, 14/09/2026] Ro: Dear OpenAI Team,