Note quite sure from the docs:

- https://openai.com/policies/data-processing-addendum 8. Term; Data Return and Deletion

- https://openai.com/policies/privacy-policy 4. Your rights (Delete your Personal Information from our records) [note: BTW "for your rights" really on purpose?]

Do you really think they would redact valuable training data?

In Europe the GDPR rights should apply, but even then, it is always a leap of faith.

Now thinking, it's probably because of European GDPR that they had to add the `is_visible` flag to `false`, "to then handle properly removing the data in a clean job depending on the user region", to see which data can be "not deleted and trained-on/sold-to-3rd-parties or hidden for GDPR eventual scrutiny").

> GDPR request response: "yes we do not have even an empty JSON object" about any ID referring to future previous or actual uniquely identifiable ID bound to any of your device or writing pattern" /paranoia "off'

They do say no data is used for training in the chat GPT UI no?

At this point assuming that a tech company deletes anything and not put it in deep storage with "very deleted" flag is just naive if they can afford the storage.

Naivety for sure… but why obviously lying?

Maybe with next sprint they'll rename the endpoints.

Highly doubt.

Even Snapchat didn’t delete the “disappearing” messages lmao.