Privacy question on image, data off

[neon cloak]

Hey, we're trying to help a user who is asking:

"I'm trying to see if when data controls are off if images you upload still get seen by other people or not. I used a stock image into DALL-E directly (ChatGPT app on Android) and you're able to directly tell it what to do as opposed to having to get a description from ChatGPT to feed into DALL-E."

They said it in response to this conversation:

"So I got a question. Sadly we can't use DALL·E3 to upload photos of ourselves and kind of have it make avatars or things like that, and I really like making avatars using DALL·E3. Is there any way or anything you all might be aware of where we could upload our photos and get an accurate description of ourselves so we could get DALL·E3 to generate an avatar as close as possible to our resemblance?"

"Being mindful of privacy and all, and I'm not sure that uploading images even of yourself is a good idea, talk about PII, but...

if you upload an image to Vision and ask it to describe it completely for use as a Dall-E-3 prompt in whatever style you wish, then check the description, fix any errors you notice, then put the description into Dall-E-3, you may get an image that is a lot like what you are hoping for."

Let's check all files related to data controls, privacy, infomation sharing retention for safety checks, and related information.

[manic sierra]

Hi @neon cloak. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:

---

Based on the information provided in the extracts, OpenAI has implemented several data controls and privacy measures to ensure the safety and privacy of user data. Here are some key points:

1. Data Controls: OpenAI provides data controls that allow users to turn off chat history and choose whether their conversations will be used to train OpenAI's models. Users also have the option to export their ChatGPT data and permanently delete their account source (https://help.openai.com/en/articles/7730893-data-controls-faq#h_4b939fe966).

2. Privacy and Safety: Creators of GPTs cannot access user conversations with their GPTs. OpenAI has automatic systems to help ensure GPTs adhere to usage policies, preventing harmful content and impersonation. Users can report concerns for further review source (https://help.openai.com/en/articles/8554407-gpts-faq#h_56e44d6ee1).

3. File Uploads: OpenAI may use content submitted to ChatGPT, DALL·E, and other services for individuals to improve model performance. This content may include files that are uploaded. However, OpenAI does not use content submitted by customers to business offerings such as the API and ChatGPT Enterprise to improve model performance source (https://help.openai.com/en/articles/8555545-file-uploads-with-gpts-and-advanced-data-analysis-in-chatgpt#h_3137baa34c).

4. Access to User Content: A limited number of authorized OpenAI personnel and trusted service providers may access user content only as needed for specific reasons such as investigating abuse or a security incident, providing support, handling legal matters, or improving model performance source (https://help.openai.com/en/articles/7039943-data-usage-for-consumer-services-faq#h_9a62c8668e).

In the context of using DALL·E3 to generate avatars from personal photos, it's important to note that while OpenAI has robust data controls and privacy measures in place, uploading personal images could potentially involve privacy risks. It's recommended to be mindful of these risks and make informed decisions about the use of personal data.

As for the question about whether uploaded images can be seen by other people when data controls are off, the extracts do not provide specific information on this. It would be best to refer to OpenAI's official Privacy Policy (https://openai.com/policies/privacy-policy) and Terms of Use (https://openai.com/policies/terms-of-use) for detailed information.

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon: