3 ms·

Everyone knows that they train on the discounted rate plans data. All the labs are upfront about this too. If you need privacy, then you are going to have to p

Everyone knows that they train on the discounted rate plans data. All the labs are upfront about this too. If you need privacy, then you are going to have to pay full price for those tokens (API). This has been true since day one. Everyone knows it, I guess though this is the first time that it has become "real".

perching_aix 1mo ago
There's literally an opt out toggle even pesky peons like me can peruse, actually.
- lima 1mo ago
  They may still train on it if you submit feedback or flag a safeguard. The terms are a bit fuzzy on this.
spruce_tips 1mo ago
what counts as discounted rate plans? if i pay for a year in advance (and get the yearly discount) and have train on my data set to off.. are you saying that is still being trained on?
- magicalhippo 1mo ago
  It's quite well explained here[1], which is linked from the Privacy section of their plan overview[2]. Basically individual accounts can opt out, while business and enterprise plans as well as API users can opt in. You'd have to take their word, but that goes for anything in life. [1]: https://help.openai.com/en/articles/5722486-how-your-data-is-used-to-improve-model-performance https://help.openai.com/en/articles/5722486-how-your-data-is... [2]: https://chatgpt.com/pricing/ https://chatgpt.com/pricing/
14u2c 1mo ago
You can also pay for their business plan, which includes data controls and starts at $50/mo (2 seats). Not exactly a high bar.
TZubiri 1mo ago
>They only fuck over the poor ones, I can pay the expensive prices so this is not a problem.
nozzlegear 1mo ago
> If you need privacy, then you are going to have to pay full price for those tokens (API). At this point, how can we even trust that they aren't accidentally training on those tokens too?
- jdm2212 1mo ago
  It'd be corporate suicide for them to be caught violating zero-data-retention commitments. But also if you're really paranoid you can just use ChatGPT on Azure or AWS, where nothing is flowing back to OpenAI at all.
  - nozzlegear 1mo ago
    > It'd be corporate suicide for them to be caught violating zero-data-retention commitments. One would think getting caught asleep at the wheel while their bots are escaping containment and hacking third parties would be corporate suicide. One would think that potentially stealing their competitors' work on the Navier-Stokes problem would be corporate suicide. Alas we live in bizarro world where there are zero consequences (maybe the opposite, in fact) for the first, and their employees meme about the second on social media.
    - WarmWash 1mo ago
      Neither of those other two examples would stop corporations from giving you money. Bold faced lying to them would. Boardrooms run businesses, not bookstore ethics clubs.
      - nozzlegear 1mo ago
        Boardrooms don't typically make mundane purchasing decisions like "which AI vendor should we use."
        jdm2212 1mo ago
        The boardroom will absolutely veto a decision like "let's give all of our proprietary data way to a company that will use it to train a competing product". That's why zero data retention exists, and why it's corporate suicide to not do it correctly.
thewhitetulip 1mo ago
Umm did you hear about the huggingface incident where they were training a model and that model started using artifactory as an internet gateway + shared wiki? Or how about the German Wiki that openai agents overtook?

perching_aix 1mo ago

There's literally an opt out toggle even pesky peons like me can peruse, actually.

lima 1mo ago
They may still train on it if you submit feedback or flag a safeguard. The terms are a bit fuzzy on this.

spruce_tips 1mo ago

what counts as discounted rate plans? if i pay for a year in advance (and get the yearly discount) and have train on my data set to off.. are you saying that is still being trained on?

magicalhippo 1mo ago
It's quite well explained here[1], which is linked from the Privacy section of their plan overview[2]. Basically individual accounts can opt out, while business and enterprise plans as well as API users can opt in. You'd have to take their word, but that goes for anything in life. [1]: https://help.openai.com/en/articles/5722486-how-your-data-is-used-to-improve-model-performance https://help.openai.com/en/articles/5722486-how-your-data-is... [2]: https://chatgpt.com/pricing/ https://chatgpt.com/pricing/

14u2c 1mo ago

You can also pay for their business plan, which includes data controls and starts at $50/mo (2 seats). Not exactly a high bar.

TZubiri 1mo ago

>They only fuck over the poor ones, I can pay the expensive prices so this is not a problem.

nozzlegear 1mo ago

> If you need privacy, then you are going to have to pay full price for those tokens (API). At this point, how can we even trust that they aren't accidentally training on those tokens too?

jdm2212 1mo ago
It'd be corporate suicide for them to be caught violating zero-data-retention commitments. But also if you're really paranoid you can just use ChatGPT on Azure or AWS, where nothing is flowing back to OpenAI at all.
- nozzlegear 1mo ago
  > It'd be corporate suicide for them to be caught violating zero-data-retention commitments. One would think getting caught asleep at the wheel while their bots are escaping containment and hacking third parties would be corporate suicide. One would think that potentially stealing their competitors' work on the Navier-Stokes problem would be corporate suicide. Alas we live in bizarro world where there are zero consequences (maybe the opposite, in fact) for the first, and their employees meme about the second on social media.
  - WarmWash 1mo ago
    Neither of those other two examples would stop corporations from giving you money. Bold faced lying to them would. Boardrooms run businesses, not bookstore ethics clubs.
    - nozzlegear 1mo ago
      Boardrooms don't typically make mundane purchasing decisions like "which AI vendor should we use."
      - jdm2212 1mo ago
        The boardroom will absolutely veto a decision like "let's give all of our proprietary data way to a company that will use it to train a competing product". That's why zero data retention exists, and why it's corporate suicide to not do it correctly.

thewhitetulip 1mo ago

Umm did you hear about the huggingface incident where they were training a model and that model started using artifactory as an internet gateway + shared wiki? Or how about the German Wiki that openai agents overtook?