OpenAI is now offering its leading conversational AI to everyone, even those who haven’t bothered creating an account. However, the experience may not be identical, and all conversations will still be used for training data unless you choose to opt out.

Starting today, chat.openai.com will no longer require you to log in, initially in select markets and gradually expanding globally. Instead, you’ll be immediately engaged in conversation with ChatGPT, utilizing the same model as logged-in users.

You can converse freely, but it’s important to note that you won’t have access to all the features available to account holders. Functions such as saving or sharing chats and using custom instructions will not be available without a persistent account.Nevertheless, you retain the option to prevent your conversations from being used for training purposes, although this action may counteract the company’s primary motivation for implementing this change.

The free version of ChatGPT will now have “slightly stricter content policies.” Essentially, this means that precautions are being taken to ensure that the generated content remains safe and appropriate for users who are not logged in.

These measures include existing safety features built into the model, such as preventing the generation of harmful content. Additionally, new safeguards are being implemented to address other types of content that may not be suitable for users who are not logged in.

To determine these measures, potential misuse scenarios were considered, taking into account the capabilities of GPT-3.5 and risk assessments conducted by the company.

So, there’s no clear indication of what exactly these stricter policies entail. It seems we’ll have to wait and see as users test out the new offering. The spokesperson mentioned that they’re open to feedback and recognize that further adjustments may be necessary.

Regarding potential abuse and misuse of the model, I inquired if there was a plan in place to address the likely influx of attempts to exploit it. Given the significant resources required for inference, especially with a refined model like GPT-3.5, it’s expected that people will push its limits. This could result in unprecedented challenges, including scenarios where platform usage leads to financial losses for individuals. It remains to be seen how such situations will be managed.