321–330 of 441 posts
Earlier quoted context omitted.
Encryption that you have the keys to won't save you from a court order
what about encryption only the users have the keys to? I'm assuming thats what parent meant
This is kind of a fundamental issue in cloud computing: it's someone else's computer, which means it's involved in someone else's legal disputes.
Chats contain way too much sensitive private data to subject them to bulk fishing expeditions
Earlier quoted context omitted.
> What is the purpose of OpenAI storing millions of private conversations Your previous ChatGPT conversations show up right in the ChatGPT interface. They have to store the private conversations to enable users to bring them up in the interface. This isn't a secretive, hidden data collection. It's a clear and obvious feature right in the product. They're fighting for the ability to not retain secret records of past c…
They could have been stored at the client, and encrypted before optionally synced back to OpenAI servers in a way that the stored chats can only be read back by the user. Signal illustrates how this is possible. OpenAI made a choice in how the feature was and is implemented.
Earlier quoted context omitted.
I get the feeling, but that's not what this is. NYTimes has produced credible evidence that OpenAI is simply stealing and republishing their content. The question they have to answer is "to what extent has this happened?" That's a question they fundamentally cannot answer without these chat logs. That's what discovery, especially in a copyright case, is about. Think about it this way. Let's say this were a book store…
> The user has no right to privacy. The same as how any internet service can be (and have been) compelled to produce private messages. The legal term is "expectation of privacy", and it does exist, albeit increasingly weakly in the US. There are exceptions to that, such as a subpoena, but that doesn't mean anyone can subpoena anything for any reason. There has to be a legal justification. It's not clear to me that su…
Earlier quoted context omitted.
Signal does End-to-end encryption, so they (Signal) can never read it. The whole point of ChatGPT conversations is so they can be read by the model on the server. Conversations are kept around because they can be picked up and continued at any point (I use this feature frequently). Additionally you can use conversations in their scheduled notification feature, where the conversation is replayed and updates are sent t…
You could definitely do this E2EE. Models should run in ephemeral containers where data is only processed in RAM. For active conversation a unique and temporary key-pair is generated. Saved chats are encrypted client side and stored encrypted server side. To resume a conversation[0], decrypt client side, establish connection to container, generate new temporary key-pair, and so on. There's more details and nuances bu…
Mullvad and E2EE Messengers do not need to process the contents of the message on their server. All they do is, passing it to another computer. It could be scrambled binary for all they care. But any AI company _has_ to read the content of the message by definition of their service.
Earlier quoted context omitted.
>NYTimes has produced credible evidence that OpenAI is simply stealing and republishing their content They shouldnt have any rights to data after its released. >That's a question they fundamentally cannot answer without these chat logs. They are causing more damage than anything chatGPT could have caused to NYT. Privacy needs to be held higher than corporate privilege. >Think about it this way. Let's say this were a…
I get that you're mad, and rightly should be for an invasion of your privacy, but the NYT would be foolish to use any of your data for anything other than this lawsuit, and to not delete it afterwards, as per their request. They can't use this data against any individual, even if they explicitly asked, "How do I hack the NYT?" The only potential issue is them finding something juicy in someone's chat, that they could…
Which is concerning since this is a news organization that's getting the data.
Let's say they do find some juicy detail and use it, then what? Nothing. It's not like you can ever fix a privacy violation. Nobody involved would get a serious punishment, like prison time, either.
Earlier quoted context omitted.
This is true but why retain deleted conversations?
ChatGPT (the app) specifically says they keep deleted conversations for up to 30 days. That's probably why.
Earlier quoted context omitted.
> That's the point of the case. No, its not. See the PDF of the actual case below. The case is largely about OpenAI training on the NY Times articles without permission. They do allege that it can reproduce their articles verbatim at times, but that's not the central allegation as it's obviously a bug and not an intentional infringement. You have to get way down to item 98 before they even allege it. https://nytco-as…
They alleged it in point 4? "Defendants have refused to recognize this protection. Powered by LLMs containing copies of Times content, Defendants’ GenAI tools can generate output that recites Times content verbatim, closely summarizes it, and mimics its expressive style, as demonstrated by scores of examples. See Exhibit J. These tools also wrongly attribute false information to The Times."
Still, that's a bug not a feature. OpenAI will just respond that its already been fixed and pay them damages of $2.50 or something to cover the few times it happened under very specific conditions.
Earlier quoted context omitted.
Signal does End-to-end encryption, so they (Signal) can never read it. The whole point of ChatGPT conversations is so they can be read by the model on the server. Conversations are kept around because they can be picked up and continued at any point (I use this feature frequently). Additionally you can use conversations in their scheduled notification feature, where the conversation is replayed and updates are sent t…
You could definitely do this E2EE. Models should run in ephemeral containers where data is only processed in RAM. For active conversation a unique and temporary key-pair is generated. Saved chats are encrypted client side and stored encrypted server side. To resume a conversation[0], decrypt client side, establish connection to container, generate new temporary key-pair, and so on. There's more details and nuances bu…