Comment by buzer
3 hours ago
There are multiple ways to use feedback. Personally I would often be fine with actual human reading my feedback, taking in account the points I made and evaluating how it should affect their feature development and future roadmap.
Now what I would expect AI companies to do is to take things which were submitted as feedback and pretty much adding to training:
"Do more of this: <copy of the whole response which was flagged as good in feedback>"
"Do less of this: <copy of the whole response which was flagged as bad in feedback>"
It's paraphrased, but the point is that they will most likely use it more-or-less as-is and thus whatever is in there will be part of the model's training set rather than someone picking up the parts from response that are important and only including them (which happens with traditional feedback).
If you have a naive loop like this you will quickly poison your dataset with e.g. thumbs downs because someone is unhappy with the latest frontend update, or teach models to not correctly refuse. I’m sure the pipeline is more sophisticated and in the middle.
Sure, but the point is more that once you submit feedback then the usual "opt out of using my data for training" no longer apply and at least that reply (and possibly whole conversation) can be included in training set in one way or another.
But we already know they're not using the transcript for training (unless you opted in t that)
They don't use your general chats it if you have opted out (note: opted out, not opted in). However if you submit feedback then whole conversation can be used.
https://help.openai.com/en/articles/5722486-how-your-data-is...
> Even if you have opted out of training, you can still choose to provide feedback to us about your interactions with our products (for instance, by selecting thumbs up or thumbs down on a model response). If you choose to provide feedback, the entire conversation associated with that feedback may be used to train our models.
https://privacy.claude.com/en/articles/7996885-how-do-you-us...
> If you explicitly report materials to us (e.g.via our thumbs up/down feedback mechanisms), or by otherwise explicitly opting in to training, then we may use those materials to train our models.
Again, that's a yes for OpenAI and a no for Anthropic.
4 replies →