OpenAI appears to be inconsistent with its privacy policy, as it still retains memory of user interactions despite opting out and subscribing to GPT Plus. I personally conducted a test to verify this claim. I composed a piece of text and submitted it to GPT-4 using my opt-out account, then asked if the text was generated by GPT. The response was: No i did not Generate it.
Subsequently, I logged into my separate GPT-3.5 free account, provided the same text, and inquired if GPT had generated it. Surprisingly, the response confirmed that GPT had Generated the text. When asked for an explanation, the AI claimed that another user had already prompted GPT with the same content, leading GPT to believe it had generated the text. This implies that GPT retained the prompt data from my opt-out account, which contradicts its privacy policy.
I experienced the same outcome after multiple attempts and am both shocked and frustrated by these findings. I am seeking clarification from the OpenAI team regarding this issue.