One of the main problems when using ChatGPT is without a doubt regarding privacy. Just a year ago, Italy blocked access to the chatbot because the service may have violated the European Union’s data protection legal framework, an action that prompted other EU regulators to consider measures along the same lines. And while it is true that OpenAI acted conscientiously and managed to reverse the blockade, it was only two months ago that we learned that the Italian regulator still doubts it.

It is true that since April of last year, OpenAI allows users to configure their ChatGPT account so that the model is not trained on their conversations, as well as clear the entire conversation history in one swipe with a chatbot. But even so, it is still advisable, strongly recommended, not to enter any personal/private/confidential data into the conversations, especially if we are not sure if we have disabled the possibility of the model being trained with them.

Of course, ChatGPT having memory isn’t necessarily a bad thing. Exactly opposite, This can be very useful in many cases., and a good proof of this is the fact that last February OpenAI added a feature designed for this very purpose. The key, of course, is that it is the user who has control over what the chatbot is able to “remember”, as well as the ability to review it and delete it if necessary.

Now there are those who seek the opposite extreme, and this is completely understandable, to the extent that alternative proposals have already appeared that are characterized by offering anonymous access, which we can also find in Copilot. And now, as we can read in TechCrunch, OpenAI is adapting to this demand and has started offering anonymous access to ChatGPT. This means that you will not need to log in to start a conversation with the chatbot.

This measure will be introduced gradually, i.e. not yet available worldwide and will also have certain restrictions regarding non-anonymous access. For example, we read that the content policies will be somewhat more restrictive as more filters will be added to the existing ones. On the other hand, logically, we will not maintain a history of conversations, and in addition, we will not have features like sharing conversations.