OpenAI has clearly stated that they will use user input for training so it's now Google's responsibility to keep their confidential information from ChatGPT inputs, assuming that OpenAI can make unintentional mistakes. What's wrong in here? And your claim is a completely different story, Google and other cloud services clearly state that they will handle user data separately so it's their liability if something goes wrong. And you're trying to insist that they will secretly use user data for whatever they want, breaking their public promises and putting their own business at severe legal risks. So tell me, why would they do that?

> We already know that state level actors are hoovering up civilian data en masse, so this is probably already happening.

Please don't put your own precious conspiracy unless you have plausible evidence. That only harms the credibility of your claim and deteriorate signal to noise ratio in this forum.

This isn't speculative. We know that the government uses Google (and others) to access pretty much all data on the internet:

https://en.wikipedia.org/wiki/PRISM

From the article:

> Internal NSA presentation slides included in the various media disclosures show that the NSA could unilaterally access data and perform "extensive, in-depth surveillance on live communications and stored information" with examples including email, video and voice chat, videos, photos, voice-over-IP chats (such as Skype), file transfers, and social networking details.

Considering their chat product is less than a year old, and has had multiple bugs that expose chats to other customers, and has been shown to not be honoring “delete” requests… I think it’s safe to say that there’s evidence that OpenAI is sufficiently negligent even if not nefarious.

Oh that’s to say nothing of the actual data being used in training. Imagine if a bing engineer can just ask a GPT for info on what google is planning and it spits out data it was trained on? Big risk.