It's hard to believe it's been six months since the initial demo of OpenAI's visual AI that we were told could identify just about anything and even solve math equations, but we may finally be getting closer to some sort of rollout.
As spotted in the code of the latest ChatGPT beta, OpenAI's app now has references to 'Live Camera' video features that would essentially add 'eyes' to its very impressive (and conversational) Advanced Voice Mode.
First seen by Android Authority, code in version 1.2024.317 reveals "Live camera functionality", "Real-time processing", "Voice mode integration" and "Visual recognition capabilities".
This would allow you to open the webcam while talking to Advanced Voice so it can give live feedback on what it can see in front of you. It is similar to Google's anticipated Project Astra with real-time visual analysis.
NEW: ChatGPT’s ‘Live Camera’ video features were just found in the code of the latest beta.Six months after their initial demo, OpenAI's visual AI could be ready for testing — and Advanced Voice Mode could soon be getting ‘eyes’The code discovered in v1.2024.317 reveals:… pic.twitter.com/yzwAExTuUANovember 20, 2024
OpenAI's 'Live Camera' could be closer than ever
The strings found in the beta version of the ChatGPT Android app suggest the Live Camera feature could arrive as part of a ChatGPT beta in the near future.
In the demos from May during OpenAI's Spring Update, the video features could recognize a dog, its actions with a ball, and more, while remembering key pieces of information like the dog's name.
Another, later demo, showed someone using ChatGPT Live Camera while touring London to have it point out details of different locations and landmarks.
Live demo of GPT-4o's vision capabilities - YouTube
While we've had Advanced Voice rollout to everyone, including in the web, things have been quiet on Live Camera in the months since the announcement.
Android Authority also says it "spotted warnings for users that advise them not to use the Live camera feature for live navigation or other decisions impacting their health or safety."
Here's hoping for news soon, as it was definitely one of the most impressive reveals we've seen from OpenAI so far.
More from Tom's Guide
- 7 prompts to create amazing images with ChatGPT
- ChatGPT vs Gemini vs Meta AI: Thanksgiving AI image generator shoot-out
- Apple could be giving Siri a ChatGPT-like upgrade — here's what we know