Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple Bans Employees From Using ChatGPT Amid Its Own AI Efforts
- Thread starter MacRumors
- Start date
- Sort by reaction score
Just like any other large corporation or consulting firm, Apple has information security policies which govern what its employees can and can't do. While you would think so, it's obvious from some of these comments that some people aren't smart enough to understand that you don't upload confidential data to a public resource like an Ai which has already been hacked. Similarly you aren't allowed to upload confidential data to public places like GitHub or forward it to your personal email.
Data is tagged in different ways, usually public, private, internal, and confidential. Most working code is considered confidential.
Last edited:
That is an interesting statement. He has presided over the growth of Apple into the tech behemoth that it is today, with a diverse portfolio of devices and services. I’d say he has that as a legacy.
That’s a big legacy to bring. As a customer and investor that legacy works for me.
I agree. It’s only going to stifle their attempts to compete with the technology and will just breed ignorance. It’s a fear based risk response and while it may seem “prudent” it’s hardly indicative of a business entity that is forward thinking an innovative.
Just keep this in mind. How exactly is Apple ensuring nothing of consequence is leaking out over any other medium of interest?
Did we ban email? World Wide Web? Guess what. Yeah. Companies actually did that for a long time and those who relented slowly but surely came around. The dinosaurs who waited too long are barely living relics at this point.
You’d think these fools in the tech sector would actually learn from the mistakes both they and their predecessors made but humans are basically kept stupid by their own fear. So sad. LOL. ChatGPT save us!!! LOL. 😂😂🤣
Having a private instance would only reduce IP leakage. And that's good; don't get me wrong.
But I bet they really want to minimize the risk of copying other's ideas - or at least the perception of copying. Whether it was deliberate or unintentional copying really doesn't matter once the lawyers get involved: Mud's slung - reputation's tarnished. $$$$
Your statement about Cook's legacy might not age well about 2 weeks from now but we'll soon see. But I also see a flaw in your logic. Apple has become 10x more valuable under Cook since he took over but that doesn't necessarily mean he's only been chasing the money. The same thing could've easily happened if Jobs had lived to run the company to present day. The real money grab right now is to proclaim partnerships or competitive solutions to OpenAI which everyone but Apple is doing. Apple has always shown restraint by not playing Wall Street's hype game under both Jobs and Cook. They have always been a traditional and rather conservative hardware design company and I don't see that has changed much under Cook. Aside from services, most of Cook's moves were already telegraphed by Jobs. They don't make big bets anymore which is the excitement that might be missing from Apple, but that's a gambler's mentality. We consumers love it because it's exciting whether Apple wins or loses because someone will always come along to offer an alternative solution for our needs. We just want it to be Apple because we know they have the best taste and usually the best implementations. But the excitement of announcements are all but gone mostly due to leaks. That is a function of scale and not poor secrecy measures. It's impossible to keep a secret from an entire industry fed by thousands of leakers across the globe. It benefits Apple but also hurts the excitement which trickles down to consumer excitement.
Maybe Cook's legacy will only be about $ but I think he has taken Apple to a place where they can begin to make huge bets on health, banking, telepresence and possibly automobiles. These are calculated bets so he's not risking it all like Jobs did with the Microsoft deal and the iPhone release. But if Cook does announce a MR a headset in June, that is proof right there that Apple is still taking some risks by bucking current trends (AI) and embracing older ones (VR) that the market has essentially given up on.
You could also look at it this way.
According to many people, Siri is objectively worse than other voice assistants like Google Assistant, and yet it hasn't stopped Apple from becoming one of the most successful companies in the world. Products like the apple watch bring Siri to your wrist front and centre, while also preventing competing voice assistants from finding any sort of traction amongst Apple users.
Meanwhile, Google has the best search engine, yet continues to pay Apple billions of dollars every year just to keep google search as the default on mobile safari. For all the market share that android enjoys, iOS consumers continue to be where the money is.
It's very well possible that other companies have a better LLM than Apple, and Apple maintains its dominance by virtue of its massive ecosystem, which in turn guarantees that competitors still have go through Apple in order to reach customers, rather than supplant Apple altogether.
It's not about who has the best tech on board, but who has the better go-to-market strategy (ie: who monetises their product best).
I 100% agree with your points.
Exactly the way I use it, instead of having to navigate stack overflow’s unpleasant answers or pages of documentation (both of which still have a real and valid use), chatgpt has been fantastic for a first quick working state (plus learning how it’s done). Plus the iterative nature is great, later on your swift class stored chat, you can ask it to “add another field to the swift data class”, or “make it conform to x or y protocol” and see what comes out.
Also great for corporate spiels and HR blabbers… it’s like having the armies of assistants that the highest higher-ups have for every communication they want to send, but for the layman in the trenches.
What I understandably see happening is that it only takes ONE person to ask the AI about Apple Car or some other highly confidential endeavor, hence they have to carpet bomb the whole thing and everybody pays the price.
But not on the “deducing secrets” powers of chatGPT… I don’t see how asking for random FooBars boilerplate on random languages can somehow result in top secret information conclusions. Millions are or will be using it that way, what makes it say that for one FooBar prompt it’s an apple car endeavor but for another person it’s a learning hobby calculator app. Unless I’m underestimating its powers.
EDIT: alright, some commenters make the point of “in aggregate on thousands of data points” then it could get far finding out what a corporation might be doing.
It does make sense but still find it far fetched? in the same way that we could theoretically predict the future because we have understanding of micro and macro physics, so we would “just” have to aggregate what all the particles and objects are doing right now and extrapolate forward.
Last edited:
Good point, I hadn’t considered that they were worrying about tainting the other way around.
And people seem to still discard the reality distortion field when the proof is right there for everyone to see.
Oh nah, I've definitely been reading documentation and learning from senior devs at work. (We've already had the talk about not wanting to fire anyone for using AI with sensitive info)
Just was curious at how meticulous you'd have to be when it came to privacy with some of these prompts potentially. (Thank you for responding to this btw)