Sarah Silverman leads class-action lawsuit against ChatGPT creator

[UngainlyTitan]

I wish everybody realized that if there's no product being sold to you on a website, your information is being sold as the product to others.

It is worse than that, even where you are buying product (say on Amazon or eBay) your data is being mind for additional profit to advertisers and the like.
As more and more of our critical social infrastructure move online and we are authenticating our connections to online services via email addresses people are at risk of becoming non persons is they ever get banned by their email provider and the like.

 

[Hussar]

But, that's not new. Back in 1950s, they noted that the 60s would be ugly. In the 60s, the 70s were going to be ugly, and so on.

There never was a golden time when the future wasn't going to be ugly.

Yeah but they were actually wrong. The sixties were economically fantastic compared to the fifties and the seventies weren’t too bad either.

You don’t get the massive closures of factories until the eighties.

Put it another way. Where was the Rust Belt in 1975?

We’re about to see the same sort of shift that the eighties saw in blue collar work hit the service industry. It’s going to be a huge upheaval.

 

[overgeeked]

I think "AI" might be the first really good, really wide-spread example of Clarke's Third Law: "Any sufficiently advanced technology is indistinguishable from magic."

So-called "AI" is not magic. It's a program. It's fed data (human artists' work). It then is iteratively programmed to produce sufficiently similar works, whether image or text. There's no "neural network" here. There's no imagination here. No dreaming. No intelligence. It's literally a string of code illegally fed on human artists' copyrighted works. "Generative AI" is a nice, corporate-speak mask for what they really are. Plagiarism machines.

 

[doctorbadwolf]

Midjourney copies images just like LLMs copy text. That’s why you end up with watermarks and logos. The images Midjourney produces only appear novel because it’s drawing from more discrete sources. It’s still a plagiarism engine. Yes, ChatGPT copies text. That’s what these programs do. Take from multiple sources, copy them, cobble them together, and spit out content. It’s just harder with text because it’s harder to model coherent language than shaded pixels to produce images.

Chatgpt is also more worrying, IMO, because it fools enough people that the general populace thinks this is actual AI, that chatGPT knows things, and that just isn’t the case in any way. It’s just a predictive algorithm with a large dataset.

Of course that’s true of art “ai” programs, as well.

 

[Reynard]

“Hey, people who know more about it please explain it to me…no, you’re wrong.”

Well, i did say I have read most about Midjourney and everything I have read talks about how the engine examines patterns of relationship between text and images, then uses those patterns to output novel images based on text input. It doesn't create collages. If that isn't true, I'd love for you to point me toward a discussion that explains it.

 

[overgeeked]

Chatgpt is also more worrying, IMO, because it fools enough people that the general populace thinks this is actual AI, that chatGPT knows things, and that just isn’t the case in any way. It’s just a predictive algorithm with a large dataset.

Of course that’s true of art “ai” programs, as well.

Exactly. Humans are easily fooled because we have evolved to see faces where there aren't any and assign motive and intelligence to inanimate objects.

I don't remember where I saw this, but I thought it was great. It's not the first AI that passes the Turing test we need to worry about, we need to worry about the first AI that intentionally fails the Turing test.

There was a great article about humans anthropomorphizing AI. I can't seem to find the one I'm looking for, but here's a similar article.

AI isn’t close to becoming sentient – the real danger lies in how easily we anthropomorphize it

To what extent will our psychological vulnerabilities shape our interactions with emerging technologies?

bigthink.com

 

[Moonmover]

That and the fact that they're completly ignorant of how AI works, are applying a double standard (how come Warhol can steal entire product designs, duchamp can present a mass-manufactured toilet as an original piece, and William S Burroughs can write all his pieces by literally cutting up preexisting articles and rearranging the words, but if a computer borrows a handful of pixels or half a dozen letters suddenly the world is ending)

The concern is that a great deal of the books the AI was trained on may have been acquired illegally.

Duchamp didn't steal that toilet. That would be a crime.

 

[overgeeked]

The concern is that a great deal of the books the AI was trained on may have been acquired illegally.

Duchamp didn't steal that toilet. That would be a crime.

Even if the books LLMs are trained on were purchased, it's still copyright infringement to create a program that spits out the text or derivative works to a third party.

 

[Umbran]

Yeah but they were actually wrong. The sixties were economically fantastic compared to the fifties and the seventies weren’t too bad either.

I am not limiting my view of what is good and bad about periods of history to high level domestic economic metrics.

 

[Sacrosanct]

When David Holz (founder of MidJourney) comes right out and says that Midjourney uses art without permission, that's a good enough source for me. Do we need a better source than literally the horse's mouth?