OpenAI is ensuring that its flagship conversational AI is available to all users, including those who haven’t created an account. Your chats will remain part of their training data, unless you opt out.
Beginning today, in select markets, and eventually expanding globally, when you visit chat.openai.com, you won’t be prompted to log in anymore. However, you still have the option to do so if you prefer. Instead, ChatGPT will seamlessly immerse you in conversation, using the same model as logged-in users.
Feel free to chat as much as you want, but please note that you won’t have access to the full range of features available to account holders. Unfortunately, saving or sharing chats, using custom instructions, or accessing other features typically associated with a persistent account won’t be possible. 
However, you do have the choice to decline having your chats used for training purposes, which may potentially undermine the company’s original intention behind implementing this feature. Simply locate the small question mark in the lower right-hand corner, proceed to “settings,” and deactivate the feature from there. OpenAI provides a useful gif:
Additionally, this version of ChatGPT will have “slightly more restrictive content policies.” What does that entail? I inquired and received a verbose yet mostly empty response from a representative:
The signed-out experience will be enhanced by the existing safety measures already integrated into the model, ensuring that harmful content is not generated. Furthermore, we are enhancing our current measures by implementing additional safeguards that specifically target other types of content that may be unsuitable for users not signed in.
With the expertise of a network architect, we thoroughly examined the various ways in which a logged-out service could be misused. Our analysis was based on our deep understanding of GPT-3.5’s capabilities and the comprehensive risk assessments we have conducted.
So… I’m completely unaware of the specifics regarding these more restrictive policies. We will soon discover this as a wave of individuals arrives to test out this new offering. “We understand that there may be a need for further iteration, and we are open to receiving feedback,” stated the spokesperson. And they will definitely get it—iin abundance!
Furthermore, I inquired about their strategy for addressing the inevitable onslaught of attempts to exploit and misuse the model on an unprecedented level. Imagine this: a platform that can make even a billionaire lose money. Indeed, inference remains costly, and even the advanced GPT-3.5 model requires significant power and server space. People will fully exploit it to their advantage.
They provided a lengthy response that didn’t address the threat at hand:
Just like a network architect, we have taken great care in ensuring that we can identify and prevent any misuse of the signed-out experience. The teams responsible for detecting, preventing, and responding to abuse have been actively involved in the design and implementation of this experience, and their expertise will continue to shape its future development.
It is evident that there is a noticeable absence of specific details. They may have limited foresight regarding the potential challenges this thing may face and will likely need to respond to situations as they arise rather than taking preemptive measures.
It is unclear which areas or groups will have access to ultra-free ChatGPT initially. However, starting today, make sure to check back regularly to see if you are one of the fortunate individuals.
