Archive · 11 Sep 2026
AI Photo Trends Raise Questions About Uploaded Image Privacy
AI apps can quickly turn your photo into an ’80s-style picture.
But uploading the photo also gives the app a copy of your image and other information.
A photo may contain details about when and where it was taken.
The app may also record information about your device and internet connection.
Different companies keep uploaded data for different amounts of time.
The new picture might disappear from social media, while the original photo stays on a company’s servers.
This does not mean the photo will definitely be leaked or misused.
It means these apps are useful but are not completely private.
AI makeover tools can collect uploaded photos, metadata, device details, IP addresses, and inferred information.
A face can become personal data when an image identifies the person shown.
ChatGPT, Gemini, and Claude have different rules for retaining chats, files, and deleted data.
AI-generated images may disappear from social media while original uploads remain on company servers.
Experts advise avoiding sensitive uploads, disabling model training, using temporary chats, and deleting old files.
- Who
- People using AI tools such as ChatGPT, Gemini, or Claude, along with the companies operating those services.
- What
- The article examines privacy and data-retention risks when users upload photos for AI image trends.
- When
- The concern is being raised as ’80s-style AI photos become popular and a recent Wired report highlights the risks.
- Where
- Uploaded images and related data may be stored on AI companies’ servers and appear in users’ social-media feeds.
- Why
- Photos can contain metadata and identifying facial information, while usage patterns may help create profiles of users.
- ChatGPT retention
- Chats remain until the user deletes them, but can stay on OpenAI’s servers for up to 30 days and longer if de-identified.
- Gemini retention
- Chat data is retained for 18 months by default.
- Claude retention
- Deleted chats are removed from Anthropic’s back end within 30 days.
- Model-training retention
- If users allow chats to help train the model, de-identified copies may be kept for up to five years.
- Photo data
- Uploaded images can include metadata such as capture time, device information, and location.
- Expert cited
- Cybersecurity expert Christina Popper of NYU Abu Dhabi discussed the layers of information AI systems collect.
- Recommended steps
- Avoid uploading identity documents or overly personal images, disable model training where possible, use temporary chats, and delete old files.