What's the point of ChatGPT team data controls?

1 point

2 years ago

I use ChatGPT for vague work-related questions sometimes (asking stuff like "how can I accomplish X technical thing using Y framework", "write me a SQL snippet that does X," "translate this from this language to that language", etc.). I avoid putting specific things that might overtrain ChatGPT on specific things about my work, it is a little annoying worrying about that but the tool is still useful.

I see that with team/enterprise licenses they say "We do not use your ChatGPT Team, ChatGPT Enterprise, or API data, inputs, and outputs for training our models." But they also say the same thing about when you opt out of training[1] or turn off chat history. So, I am confused about whether you actually get any extra data protection out of the team/enterprise packages vs. just opting out/turning off chat history... Can anyone explain if there's a real privacy/training/data ownership benefit there?

[1]https://help.openai.com/en/articles/5722486-how-your-data-is-used-to-improve-model-performance