ChatGPT Mac App Stored User Chats in Plain Text Prior to Latest Update

[ApAx]

So if the app would have been published to the App Store, it would have followed Apple’s security policies and this risk to the end user would have been prevented.

EU: “But that’s just not fair. iPhone must be as vulnerable as computers in the name of fairness! Boo hoo wah!”

 

[ApplesAreSweet&Sour]

To be fair, ChatGPT is a useful tool:

View attachment 2394388

It doesn't necessarily mean that ChatGPT has hired him to spy on users. But it also doesn't mean they won't.

At the very least, it seems like OpenAI is using him as some kind of figure head for conveying how determined they are about making AI safe and secure, to ensure the public that it's going to "benefit all of humanity", as they like to phrase it.

Who knows.

 

[6787872]

if you don't have filevault and someone has physical access to your computer, you're already screwed. way worse than your chatgpt conversations on there...

 

[Mac Fly (film)]

You’re right. AI like ChatGPT obviously have leaning bias in 1 direction that is influenced by the creators … and look who they are.

All people have to do is ask ChatGPT what and how to think about things and voila, ideology lectures 101. AI like Google’s is already pretending to rewrite historical facts.

Someone was trying to argue with me months ago on X about how AI cannot be biased. Was like talking to the wall. I left them off to talk nonsense.

 

[Mac Fly (film)]

It doesn't necessarily mean that ChatGPT has hired him to spy on users. But it also doesn't mean they won't.

I think they hired him to bring the drinks around to the tables.

 

[trip1ex]

this is the rule of thumb for anything you do on your computing devices.

 

[ApplesAreSweet&Sour]

So if the app would have been published to the App Store, it would have followed Apple’s security policies and this risk to the end user would have been prevented.

EU: “But that’s just not fair. iPhone must be as vulnerable as computers in the name of fairness! Boo hoo wah!”

That's some hyperbole if I ever heard it:

All the EU is doing is giving consumers full control of their iPhones and iPads allowing them to install apps from outside its App Stores without having to jailbreak their iPhones and void warranty, etc.

EU consumers and developers who want to stick with the convenience and safety of Apple's App Stores can do so under the same terms and with the same safety as before the DMA kicks in.

It's just more options.

But from what I gather, it's easier for Americans to act all obtuse about the DMA instead of actually comprehending what it's going to do.

When you own a house or a car, you have the right to do anything you want to it, let anyone repair it, install anything you like. It's your proporty, you have the freedom to destroy it, sell it, fix it, paint it green, whatever.

The company that built the house or the car for you cannot dictate how you can repair it, who can repair it, if third parties have to pay them fees for fixing your proporty. You are free to do it yourself or hire any other party.

It's your iPhone, you should have those same freedoms for it.

And all the shenanigans OpenAI throws at its users with most certainty shouldn't dictate the future of iOS or MacOS.

Arguing for trapping everyone inside sandboxed platforms with huge, expensive guardrails because a few people want to make some terrible decisions with their Macs, iPhones, and iPads is a very ignorant take.

 

[Speed38]

If someone is sufficiently idiotic/stupid to ask ChatGPT how to think about things, they are already a lost soul, no? "Stupid is as stupid does", said Forrest Gump, right?

 

[Speed38]

Is there any difference between downloading the ChatGPT.app to one's iPhone and/or going to the URL on a Mac, using Safari, and choosing to save the page to the Dock???

 

[IceCool]

That's some hyperbole if I ever heard it:

All the EU is doing is giving consumers full control of their iPhones and iPads allowing them to install apps from outside its App Stores without having to jailbreak their iPhones and void warranty, etc.

EU consumers and developers who want to stick with the convenience and safety of Apple's App Stores can do so under the same terms and with the same safety as before the DMA kicks in.

It's just more options.

But from what I gather, it's easier for Americans to act all obtuse about the DMA instead of actually comprehending what it's going to do.

When you own a house or a car, you have the right to do anything you want to it, let anyone repair it, install anything you like. It's your proporty, you have the freedom to destroy it, sell it, fix it, paint it green, whatever.

The company that built the house or the car for you cannot dictate how you can repair it, who can repair it, if third parties have to pay them fees for fixing your proporty. You are free to do it yourself or hire any other party.

It's your iPhone, you should have those same freedoms for it.

And all the shenanigans OpenAI throws at its users with most certainty shouldn't dictate the future of iOS or MacOS.

Arguing for trapping everyone inside sandboxed platforms with huge, expensive guardrails because a few people want to make some terrible decisions with their Macs, iPhones, and iPads is a very ignorant take.

Not necessarily true. You buy the hardware, the software is Apple's. Should I be able to do whatever I want with my car's built in OS? Guess they're an anticompetitive monopoly because I can't!

The DMA is wrong in that it's forcing a private company into opening up. The choice is already there, buy an Android if that's what you want so desperately. Forcing Apple to open up kills that choice.

And before I get the "But you can just turn it off!" rhetoric; this is also not true. Like on macOS and Windows, what about the developers that release the apps you want or - worse - need exclusively on their own "stores" or some other third party's? Well, look at the article above. Your choice of a walled garden is now completely meaningless when this happens.

I've never been against sideloading or freedom in iOS, as a previously avid jailbreaker, but the EU forcing them into it is a huge turn off.

 

[randywolf244]

Seeing how upset people are about this makes me really wonder what they're talking to chatGPT about.

 

[nortonandreev]

Who cares, as if the chats are not human-monitored anyways and/or used for training purposes…

 

[ifxf]

I don’t care, it is only stored on your computer. if you access the chats you can see bookmarks and browsing history.

 

[CalMin]

I mean, was there any real expectation of privacy here in the first place?

 

[!!!]

Am I the only one confused by this? Almost nothing aside from passwords is stored encrypted on your device, and pretty much any application can view any file on your device. If you don't trust the program not to steal data, then you shouldn't be running it at all.

Then there's this weird idea that chats with ChatGPT need to be private? They're literally all being stored on MS servers so they can further train their AI models.

 

[Ashenoa]

Someone was trying to argue with me months ago on X about how AI cannot be biased. Was like talking to the wall. I've left off to them talk nonsense.

So, what is not biased? Which media? Which person? Which service?

 

[NT1440]

It doesn't necessarily mean that ChatGPT has hired him to spy on users. But it also doesn't mean they won't.

At the very least, it seems like OpenAI is using him as some kind of figure head for conveying how determined they are about making AI safe and secure, to ensure the public that it's going to "benefit all of humanity", as they like to phrase it.

Who knows.

Lol Sam Altman gave a talk to Booze Allen Hamilton in 2016 alongside George Tenant (former CIA director) regarding AI and national security.

William Hurd, “former CIA”, had been on the board for a couple years.

Sam Altman, despite not being anything other than a conman and running several tech companies into the ground, has somehow ascended the corporate ranks of Silicon Valley every step of the way. He’s an asset.

Silicon Valley was built on the collaboration of tech and intelligence agencies. Nearly every major technology and company has spawned from DARPA projects (Google, Facebook, etc).

Let’s not be naive here folks.

 

[WarmWinterHat]

if you don't have filevault and someone has physical access to your computer, you're already screwed. way worse than your chatgpt conversations on there...

Unless you're specifically turning off FileVault, it's on, since it's on by default.

I'm not fan of openAI, and want nothing do with their products (or anything with con-artist Sam Altman), but I don't see a problem with this. It's on device...none of my office documents are encrypted beyond my Mac's encryption. My photos aren't, my email isn't, etc.

Why the big fuss?

 

[duervo]

Aside from being a massive security risk, this so-called Artificial Intelligence is nothing more than a disguised coup to manipulate public opinion, censor free thought by eliminating what they consider to be opposed to their desired schematic and also ultimately to insidiously monitor everybody's activity. Of course they will profess otherwise and efforts will be seen to create privacy, but as we see in these initial stages, first with Microsoft and now others including this, their prime intentions are not security based, our privacy is not something that they give a fig for. Otherwise, how could Microsoft even dream of putting into public space such a massive data collection facility as was their Copilot v1.0?

That makes sense. Those BBQ rub recipes I had ChatGPT create are clearly a gateway to world domination. It started with the rub, and then I asked it what side dishes would go well with the rub on pulled pork.

Let me tell you, those side dish suggestions have got my shields up more than anything else. Cornbread with dried cranberries in it?! Are you kidding me?! We’re all doomed.

 

[someshwaradas]

It seems like I hear that excuse a lot from different companies that screw up and are caught .

Something done deliberately is not a “screw up.”

 

[pesc]

Am I the only one confused by this? Almost nothing aside from passwords is stored encrypted on your device, and pretty much any application can view any file on your device. If you don't trust the program not to steal data, then you shouldn't be running it at all.

Then there's this weird idea that chats with ChatGPT need to be private? They're literally all being stored on MS servers so they can further train their AI models.

An app installed from the MacOS App Store runs in a sandbox and can not access any file on the device, unless the user actively opens the file.

While ChatGPT chats are stored in MS servers, they can not be connected to a specific user identity if Apple Intelligence was used. Not by OpenAI or MS. And chats should certainly not be accessible by other entities either.

 

[NT1440]

That makes sense. Those BBQ rub recipes I had ChatGPT create are clearly a gateway to world domination. It started with the rub, and then I asked it what side dishes would go well with the rub on pulled pork.

Let me tell you, those side dish suggestions have got my shields up more than anything else. Cornbread with dried cranberries in it?! Are you kidding me?! We’re all doomed.

The public is being conditioned to view and use AI LLM’s as the new way to search, regardless of the fact that LLMs don’t “know” anything.

In a few years time when most “reporting” is just AI, you’ve fundamentally altered the public at large’s perception of reality.

No one is claiming your bbq recipe is nefarious, no need to be obtuse.

 

[citysnaps]

Holy smokes... You mean there isn't 100% perfection 100% of the time when humans are involved?

 

[Mac Fly (film)]

So, what is not biased? Which media? Which person? Which service?

You just proved my point; they argued it wasn't biased.

For models which are known to be biased (all of them), as AI becomes increasingly powerful, they conceivably have the potential to perform serious destruction. This is why AI should be overseen by regulatory bodies with a broad variety of viewpoints and all discussion and debate should be held in the open for all to see. An additional possibility should be for the user to choose how they would like to receive their response, so they would know themselves they are the ones using their own biases right in the settings of the app, and they could switch between different models of bias.

 

[ifxf]

You just proved my point; they argued it wasn't based.

For models which are known to be biased (all of them), as AI becomes increasingly powerful, arguably have the potential to perform serious destruction. This is why AI should be overseen by regulatory bodies with a broad variety of viewpoints and all discussion and debate should be held in the open for all to see. An additional possibility should be for the user to choose how they would like to receive their response, so they would know themselves they are the ones using their own biases right in the settings of the app, and they could switch between different models of bias.

Companies and governments already can impact lives with arbitrary decisions. Just look at the government terrorist list, you can be put on it due to a similar name or because someone checked the wrong box. Once on, they don’t tell you are it is almost impossible to get off.