OpenAI introduced ChatGPT to the public in late 2022 without any privacy features, allowing data in prompts to be used to train the AI. After facing pressure from regulators, OpenAI implemented its first privacy controls, but they were not user-friendly. Users had to choose between allowing their data to be used for training or turning off all history.
In November, users were given the option to opt out of training the model while keeping their chat history, but finding this feature was not straightforward. However, in late April 2024, OpenAI quietly upgraded ChatGPT’s privacy settings, allowing users to retain their chat history and prevent data usage for training purposes.
The new privacy features were not formally announced by OpenAI and were easily overlooked by users, especially with the recent memory upgrade for ChatGPT. The company stated that the new privacy controls apply to both Free and Plus users, allowing them to access chat history regardless of training opt-in status.
