ChatGPT+ Privacy Update
Recently, I wrote a short essay about an unpleasant discovery I made with ChatGPT’s dark pattern whereby, if you want to turn off “train your models using my chats,” you need to disable one of ChatGPT’s nicest features, chat history. Here’s the dialog in question:
The update is that there may be a workaround. On the Data Controls FAQ, buried two screen-scrolls down, there’s a section that says:
What if I want to keep my history on but disable model training?
We are working on a new offering called ChatGPT Business that will opt end-users out of model training by default. In the meantime, you can opt out from our use of your data to improve our services by filling out this form. Once you submit the form, new conversations will not be used to train our models.
The takeaways:
- OpenAI seems to understand that business users will want chat history, but definitely not have sensitive data shared in chats be used to train OpenAI’s models; hence they are working on a “ChatGPT Business” offering.
- There appears to be a workaround today by filling out the form. Ironically, the form is a … Google Form.
We’ll have to wait and see whether OpenAI will close this loophole for non-business users once they launch ChatGPT Business. This announcement hints that they might.
In any case, OpenAI still gets four weasels.
