ChatGPT’s 1980s photo trend: What happens to your picture after you upload it? | Today’s news
ChatGPT’s latest viral trend encourages users to upload their selfies and turn them into 1980s-inspired retro photos. The results may look fun and nostalgic, but before uploading a personal photo, users should understand what happens to the image and the data associated with it.
A photo of a person’s face can constitute personal information because it can help identify them. An image may also contain metadata, including information about when it was taken, the device used and, depending on the file, its location. So uploading a photo means placing another copy of that image on the AI service, subject to its data retention and privacy rules, according to WIRED Middle East.
What happens when you upload a photo to ChatGPT?
When a photo is uploaded to ChatGPT, it becomes part of the interaction and may be stored along with other conversation data. WIRED Middle East notes that ChatGPT conversations are stored in a user’s account until deleted, while some data may remain in OpenAI’s systems for up to 30 days after deletion. Certain information may also be retained longer for legal or security reasons or if it has been de-identified and separated from a user account.
This does not mean that every uploaded photo will automatically be used to train the AI model. Data handling depends on the product and user settings. OpenAI says its business products and APIs do not use customer data for modeling by default.
Your photos are not the only data
The issue of privacy goes beyond selfies. Artificial Intelligence Services may process information provided directly by users, information derived from their inputs, session metadata, and files or responses that are stored.
According to WIRED Middle East, the Services may also collect basic metadata such as IP address, device type, operating system and general geographic location. Repeated interactions can potentially reveal patterns about a user’s interests, habits, or work, even if their name is not explicitly mentioned.
Should you upload your photo?
A viral trend alone does not mean that uploading a selfie will result in the image being stolen or turned into a deepfake. The larger consideration is that users are giving another company a copy of a personal image and should understand how that service stores and processes it.
Users should avoid uploading photos containing unnecessary sensitive information. They should also be careful about uploading images of other people without their permission, especially photos of children. According to media reports, one should disable model training where possible by using temporary or incognito modes, deleting chats and uploads that are no longer needed, and controlling the services the AI assistant can access.
What does OpenAI say about business data?
OpenAI says that by default, data submitted through ChatGPT Enterprise, ChatGPT Business, ChatGPT Edu, ChatGPT for Healthcare, ChatGPT for Teachers and their APIs is not used to train or improve their models. The company also says that such business data is encrypted both at rest and in transit, and that authorized organizations can configure data retention controls.
However, for people connecting to the viral 1980s photo trend through a personal ChatGPT account, the applicable privacy and data control settings are different from those that apply to commercial OpenAI products.
Before you upload that perfect retro selfie, check your privacy settings and think about what information the image reveals. The 1980s effect may disappear from social media within days, but a digital copy of a photo may remain subject to the service’s storage and deletion policies for a longer period of time.
