Chatgpt It is one of the most popular products on the Internet, with an estimated active user base 400 million per month Last February, according to Demandsage data. Users talk with her on all kinds of topics and can probably facilitate much without thinking about it, Sensitive information. Especially if the consultations are medical, legal or financial.
In this month of April, which is that of the Cybersecurity awareness In Europe, the Organization of consumers and users He has put his attention on the privacy offered OpenAI nor can they eventually expose themselves in a filtration, which deactivate a specific option of the Ia.
The concept of training of an AI refers to the data set with which it has been fed to be able to perform the tasks that are requested. By default, OpenAi uses user data, which includes the conversations they hold and interactions with AI, to train their models. If you have ever given a personal medical report to know your opinion, documents of your work to organize the information or have asked you for assistance with your finances, all that information is also stored and used by the company’s AI models.
This situation affects so much Free accounts and platform users. Only subscriptions Teams and Enterprise They do not use user data for models training by default, but accounts do Free, Plus and Pro. Openai business environments are the only ones who ensure, by contract, that the data will not be used to train the AI.
🌐 Does chatgpt use users to train?
📢 We recognize the option to share data. We tell you how to do it 👇
ℹ️ We remember that since February there are 7️⃣ practices prohibited by the regulations on the EuropeAhttps: //t.co/8r2xt6k18u Union.– OCU (@Consumnores) April 4, 2025
“Improve the model for everyone”: the chatgpt option that OCU recommends deactivating
The OCU remembers that OpenAI allows chatgpt users exclude your training data from AI modelsan option that is available following these steps:
- Open the app or access the website Chatgpt.
- In the app, play in the upper left corner menu with 2 horizontal lines And then in your Profile image. On the web, click on the profile image and then in Configuration.
- Select Data controls and deactivate the slider Improve the model for everyone.
With this configuration, you keep the history of conversations but your data will not be part of the training of language models, thus protecting your privacy.
The OCU also recommends, for the greater privacy of the user, to use the function of Temporary chat with which the conversations are not stored (they remain 30 days on the servers before being definitely eliminated) or used for AI training.
In the app, you have it available on the icon of fragmented circle in the upper right corner of the screen. In the web version, you just have to click on the button Temporary Next to the profile image.