OpenAI introduced Advance Voices in which they incorporated a little bit of friendliness to ChatGPT so that one could converse with the AI. However, there was a promised vision capability that OpenAI demonstrated during the GPT-4o announcement in May of 2024. That feature does not seem to be developed yet, but it looks like OpenAI has been striving towards that, and now it may be rolled out to a broader extent.
Based on multiple code strings, it appears that the two features highlighted in the live vision, more specifically in the Advanced Voice Mode, may be preparing for a broader beta release as well. These strings were unveiled in the latest ChatGPT v1.2024.317 beta release. When the GPT-4o was announced, OpenAI presented the Advanced Voice Mode with a vision demonstration.
While in the demo, the ChatGPT captures the subject in the camera, recalls its name, identifies the ball, and has knowledge that the ball is owned by the dog. The demo was quite good for an AI assistant, given that the assistant was almost entirely devoid of any information the user had to enter to get those responses.
ChatGPT Live Video Feature
Now, going by the strings found in the new ChatGPT beta version of the app, the feature might be named “Live Camera” once released to beta users. There are some strings (via Android Authority) to discourage users from using the Live Camera feature in ChatGPT for live navigation or decision-making that may impact their health or safety.
Notably, the codes were seen in the GPT beta update, which suggests that the feature is lined up for release soon and may be extended to ChatGPT Plus and other premium users of the AI interface.