A user asks OpenAI for a face-scan option to keep ChatGPT images consistent
A community forum post proposes an opt-in face scan so ChatGPT could keep a person's likeness steady across generated images. It is a request, not an OpenAI feature.
The short version
- The idea was posted on community.openai.com as a proposal, not as an OpenAI launch
- Users would turn their head, show both profiles and make expressions while the camera records
- ChatGPT would save a visual representation for later image generation and editing
- A reference photo upload would be an alternative, and users could skip both
- The post asks for explicit permission and full control to update or delete the data
Why it matters
The post shows a real frustration: AI-made images of the same person often drift from one picture to the next. It also shows the trade-off such a fix would bring, since better likeness would mean handing over a detailed record of your face. Nothing here suggests OpenAI is building it.
What the post actually proposes
As community.openai.com showed, a user posted a proposal for an optional feature called "ChatGPT Visual Memory 360°". It would let people record a short facial scan with their device's camera.
During the scan, the user would follow simple prompts: turn the head, show both profiles, look up and down, and make different expressions. ChatGPT would study the face from those angles and store a visual representation for later use.
The problem it tries to solve
The author wants image generation and editing to keep a person's face recognizable. The post names the hard cases: side profiles, different camera angles, varied lighting and changing expressions.
In plain terms, the writer believes a stored multi-angle reference would let ChatGPT hold on to "who you are" better than a single photo does. The post calls this more accurate, convenient and personalized.
Consent and control
The proposal puts choice up front. The scan would only run with explicit permission, and users could instead upload a reference photo or use neither option.
The author also asks for complete control over the saved data, including updating or deleting it. The post does not explain how the data would be stored beyond saying it would be kept securely, so those details are left open.
What this is, and is not
This is a user's suggestion on a public forum. Our source contains no reply from OpenAI, no timeline and no sign that the company is testing anything similar.
The title alone could be read as a memory feature already in the works, which is why the distinction matters: a wish on a forum is not a product.
What to watch
- Whether OpenAI responds to the post or comments on facial reference features
- How any such feature would store, update and delete face data if it were built
Try it
If you want consistent faces in generated images today, upload the same clear reference photo each time and say which features must stay the same.
Sources and links
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.
Listen to the AI Wire headlines
Listen to the AI WireListen to our daily AI shows every morning: on your commute, in the shower, wherever you are
Listen to the daily showsFirst reported by community.openai.com · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
