إفصاح المعالجة بالذكاء الاصطناعي

AI Processing Disclosure

Effective October 17, 2026 · Last updated October 17, 2026

This disclosure explains where BrainDV uses machine learning or artificial intelligence (AI), what each use does with your photo, and where BrainDV does not use AI. It supplements the Privacy Policy.

1. Summary

2. The AI camera assistant (on-device)

The camera assistant uses a machine learning face detection model (Google ML Kit) that runs on your phone. On the live camera screen and on the photo you take or import, it finds the face and returns its position and size, landmarks such as the eyes and chin, head angles and an estimate of whether the eyes are open. The App uses these results to:

• guide you with distance, centering and head position while you take the photo;
• count faces, and check head tilt and framing;
• position the crop so the head size and eye height match the selected photo type;
• show an advisory warning if the eyes may be closed.

The results are kept in memory for the current photo only. They are never stored, never sent to us or to anyone else, and never used to identify or recognize a person, to verify identity, or to infer age, gender, ethnicity, emotion or health. No face template is created. ML Kit does not send your photos to Google; Google states that it may receive technical metrics about the library's performance and use.

3. The checks that decide pass or fail

The checks that can block an export are deterministic: fixed rules applied to measurable values such as pixel dimensions, file size, the number of faces and the position of the face in the frame. They are derived from published U.S. requirements, listed with their sources and retrieval dates on each requirements guide. Where a property cannot be measured reliably, the App shows a warning or "unknown" instead of a fail. Final acceptance is determined by the relevant authority, not by BrainDV or any AI system.

4. No generative editing

BrainDV does not use generative AI to edit photos. It does not retouch faces, change expressions, replace or synthesize backgrounds, or add pixels that were not in your photo. If a photo cannot be cropped to the required framing using its own pixels, the App asks for a retake.

5. Optional AI advice (OpenAI)

BrainDV includes an optional feature that asks an AI model from OpenAI, L.L.C. (United States) for short written tips about a photo. It is designed to work as follows:

• Off by default. The App works fully without it. It is turned on only in Settings ("AI advisory feedback (OpenAI)"), through a consent screen that explains what is sent, who receives it and how long it is kept. Nothing is sent unless you agree.
• What is sent. A JPEG copy of the one photo you choose to check (at most 1.5 MB), with fixed instructions from us. The model receives it at OpenAI's low-detail image setting. We do not send your account ID, email address, other photos, purchase data or location to OpenAI.
• Who receives it. The App sends the photo to the BrainDV server, which forwards it to OpenAI's API and returns the answer to you.
• What the model is told. To comment only on photographic conditions (lighting, shadows, glare on glasses, head position, expression, focus), not to judge identity, age, ethnicity, gender or any other personal attribute, and not to say whether a photo will be accepted. The answer is limited to five short notes.
• Retention by us. Our server holds the image in memory only while the request runs and never writes it to our database, file storage or logs. We do not keep the advice text. We keep a usage record for each request (time, outcome, image size in bytes, token counts, estimated cost and response time) until you delete your account, to enforce usage limits and control costs.
• Retention by OpenAI. We send requests with OpenAI's option to store responses turned off. OpenAI processes the request under its API terms and policies, which govern any retention on its side.
• Advisory only. The advice is generated by an AI model and may be wrong. It is never part of the checks, never changes a check result and never changes your photo.
• Limits. Requests are limited per hour and by a daily budget, and we can switch the feature off for everyone at any time. When it is unavailable, the rest of the App is unaffected.

Current status: in the current version of the App, AI advice is not active. If you turn it on, the App saves your choice (on your device, and on our server if you have an account) but does not send any photo.

How to turn it off: switch off "AI advisory feedback (OpenAI)" in Settings. This withdraws your consent on the device and, if you have an account, on our server, and no further photos are sent. "Delete local data" also resets the choice on your device, and deleting your account deletes the consent record and usage records on our server.

6. Limitations

Face detection can be less accurate with unusual lighting, strong shadows, head coverings, glasses, several people in the frame or low image quality. The App reports uncertainty as a warning or "unknown" instead of guessing. Please check the published requirements and your final file yourself.

7. Contact

Questions about AI processing: BuroBrain LLC, email [email protected].