Is It Safe to Upload Your Photos to Gemini or ChatGPT?

Gemini: what Google says about your chats and photos

Google's Gemini Apps Privacy Hub centres on a setting called Keep Activity. When it is on, your chats and uploads are saved in your Gemini Apps Activity. The hub says activity is auto-deleted after 18 months by default, and you can change that to 3 months, 36 months, or turn auto-delete off. With Keep Activity on, Google says it uses your activity to provide, develop and improve its services, including training generative AI models.

When Keep Activity is off, the hub says future chats are not saved to your activity and are not used to train Google's AI models unless you choose to send feedback. Even then, those chats, and temporary chats, are kept with your account for 72 hours so Gemini can respond and so Google can keep the service safe. Google also says temporary chats are not used to train its AI models.

On images, the hub says that when you add an image to a prompt, Gemini uses Google Lens technology to understand what is in the image, and Google uses this information like any other prompt.

• Human review: the hub says a subset of chats is reviewed by human reviewers, including service providers, to improve quality and safety. Chats are disconnected from your account before review.
• How long reviewed chats stay: the hub says reviewed chats are kept for up to three years, and deleting your activity does not delete them because they are no longer linked to your account.
• Google's own advice: "Please don't enter confidential information that you wouldn't want a reviewer to see."
• Stopping review of future chats: Google says to turn off Keep Activity.
• Deleting: you can review and delete activity in Gemini Apps Activity and delete individual chats at any time.

ChatGPT: what OpenAI says about your chats and photos

OpenAI's help page on model improvement says that for its services for individuals (Free, Plus, Pro), it may use your content to train its models. You can opt out by turning off Improve the model for everyone under Settings, then Data controls, or by choosing Do not train on my content in OpenAI's Privacy Portal. After you opt out, OpenAI says it will not use your new conversations to improve its models. OpenAI also says it takes steps to reduce the amount of personal information in training datasets.

There is one catch worth knowing. The same page says that if you give feedback, for example a thumbs up or thumbs down on a response, the entire conversation linked to that feedback may be used to train models even if you have opted out.

• Default setting: the data controls page says data sharing is on by default for personal Free, Plus and Pro accounts, and turning it off does not delete your saved chats.
• Temporary chat: OpenAI's help page says temporary chats stay out of your history, are not used to improve its models while they remain temporary, and a copy may be kept for up to 30 days for safety purposes.
• Deleting chats: the retention page says a deleted chat is removed from your view at once and scheduled for permanent deletion from OpenAI's systems within 30 days, unless it was already de-identified and disassociated from your account or OpenAI must keep it longer for security or legal reasons. Deleted chats cannot be restored.
• Saved files: the same page says deleting a chat does not delete a file saved separately in Library or Space, so delete those too if you want them gone.
• Images: the Images help page says images you create are saved under Images, and you delete one by deleting the conversation where it was made.

The OpenAI pages we used did not describe consumer human review in the same detail as Google's hub, so we have not made a claim about it. Check OpenAI's privacy policy and Privacy Portal for the current position.

Age limits

OpenAI's Images help page also says that for a linked teen account, a parent or guardian may be able to control whether the teen can create or edit images, and teens may see a reminder before some uploads encouraging them not to share sensitive personal information.

Children's photos: a careful approach

Baby and child edits are popular in India, from first-birthday themes to Janmashtami and Christmas looks. Children cannot agree to how their images are used, so parents carry that responsibility. A few habits keep risk low without giving up the fun.

• Use a photo that shows only your child, with no school uniform logo, house number, vehicle number plate or location sign in the frame.
• Keep the edit age-appropriate. Do not make a child look older, and avoid outfits or poses you would not post of a real photo.
• Consider using a temporary chat or turning off activity saving before uploading, based on the settings described above.
• Delete the chat when you have saved the image you want.
• Share the result in a private family group rather than on a public profile, and do not add the child's full name or school.

Our baby photoshoot prompts and birthday prompts follow these principles.

Practical checklist before you upload

1. Decide whether you are comfortable with this photo being stored for a while and, in some cases, seen by a reviewer. If not, do not upload it.
2. Open the settings first. In Gemini, check Keep Activity and your auto-delete period. In ChatGPT, check Improve the model for everyone in Data controls.
3. For one-off edits, consider a temporary chat in either app.
4. Crop out anything private before uploading: documents, addresses, ID cards, screens, number plates.
5. Upload only photos of yourself or of people who have agreed. Google's hub asks users to get permission before using other people's faces or voices in Gemini.
6. Avoid feedback buttons on chats you want kept out of training in ChatGPT, since OpenAI says feedback can bring the whole conversation into training.
7. Delete the chat, and any separately saved files, when you are done.

Edited images carry provenance signals. Google's watermark help page says all media created or edited with Gemini Apps carries an invisible SynthID watermark and Content Credentials, and OpenAI's provenance page says ChatGPT images carry Content Credentials and SynthID. These mark the image as AI-made; they are not a privacy leak of your original photo. Our guides on editing with Gemini and editing with ChatGPT explain them further.