Beep boop - this is a robot. A new show has been posted to TWiT…
What are your thoughts about today’s show? We’d love to hear from you!
1 Like
I keep saying this, but none of the guests ever seem to mention it: if the authors, artists or “norms” have freely put their works on the internet, they should be fair game for AI.
If the words, images, sound or videos have been put up against the originator’s will, I.e. pirated or revenge porn or such stuff should be off limits.
If a book or image is in copyright, it should only go into the model with permission and recompense.
“It is on the internet,” should not be enough of an excuse to including it in a model, the companies should have a duty to ensure everything is legal and above board, BEFORE it goes into the model, plus the ability to have it removed, if there are any legal reasons, why it shouldn’t be there.
5 Likes
No I don’t agree at all. Just because you have my work, doesn’t mean you can do whatever you want with it. Have you not heard of plagiarism? Because, at this point, no one understands EXACTLY how we get the results from AI we do (or else we could prevent “hallucinations”,) we can’t be sure it’s not partially plagiarizing the content. There definitely needs to be limits and rules set that prevent an AI from basically “stealing ownership” of content that was made available to them in a less restricted license (buying a book doesn’t give you the right to claim it’s your own work, for example.)
2 Likes
If you put something on the public Internet, on your own site, do you forgo any claim to your ownership of those words? I don’t think so. I think that you have a reasonable expectation that those words will not be stolen and reused.
The drawback to my previous argument is that almost nobody posts in a forum where the TOS doesn’t claim to own your words, likeness, and future offspring, so I don’t really know what that means for the argument.
I agree with everything else @big_D said though
2 Likes
If you have published it on the Internet, you have pretty much given up all rights on what is done with it already - today, click farms will scrape content from news sites and forums and re-publish it to get advertising content. You might still have copyright over it, in theory, but you would be launching a dozen new lawsuits in countries with little or no copyright enforcement, if you really wanted to keep it yours.
Not in the world we currently live in, although you can protect it from LLMs, Google has come up with an AI equivalent of spiders.txt, where you can specify which parts of the site can be scraped and which must be ignored.
1 Like
If there’s a way to filter content found on (known) pirated content hosts, it should be implemented. I think scraping public works TO SUMMARIZE is totally fine. Then there’s derivatives. AI summarizes Rod Pyle’s review of Critical Mass by Daniel Suarez, should Suarez want to sue the AI? Pyle legally had the book, read it, publicly reviewed it on the show. AI displays this in its results. Nothing wrong with that instance (to me).
2 Likes
That is the thing, if it used the review to generate its content on the book, fine. If it downloaded a pirated copy of the book, not okay, in my book - same with images.
I think the only one doing AI sensibly at the moment, and it pains me to say it, is Adobe (I still bear the wounds of Flash and Adobe Reader emergency patching over years 
)
1 Like
Lol! Ah you know Flash was just awesome 
(jk)
1 Like
Can we talk about the sheer irony of the Google case? FTC sees Apple and Google making a deal about defaults and goes after Google. At least you can switch from Google on iPhone.
Meanwhile, Apple’s entire business model is around bundling their own suite of products into the OS so that it’s not even possible to switch. Is there even an option in ios to switch away from iCloud? How isn’t the FTC going after them?
The pro-Apple bias in the US is infuriating.
1 Like
Apple is definitely under the microscope both in the US and EU for its App Store practices. I think the case against buying “product placement” on an iPhone is weak, stronger against the practice on Android. But the real case against Google is as a buyer & seller of advertising. That’s the one that’s going to force divestiture.
1 Like
I appreciate that there is increased scrutiny, but I’ll wait until I see it. I get the feeling we will wait for the EU to make some decision.
What always strikes me as strange in the Google case is that they make it about search. I agree with you, advertising is the weakness. But search?? Search isn’t even a product, it’s a feature! There is no “search market” because nobody pays for it or even thinks about alternative providers. There’s just search usage which leads to ads… so make it about the ads.
Buying feature placement is the weirdest thing to make a case about. Mozilla would shrivel up without Google’s product placement fee, making Chrome more dominant. Who wants that?
In a world where nobody is charging for the features they offer, how can it be possible to find a company who isn’t using their market dominance in that non-revenue-generating feature to claw value for themselves? Unless you’re Apple and you just play by the old Standard Oil rulebook.
Who cares about paying to get an app featured when many OS-level services in both ecosystems of phones have no equivalent alternative? Good luck creating a root-level backup on any mobile phone. I bet backblaze would love to offer that service if they could even try. And I listen to WW so I know all about Microsoft’s shenanigans.
Jeez, I need a beverage.
Especially since authors, artists, and other creators “borrow” concepts from other peoples’ work. Correct me if you think I’m wrong but I’m struggling to see the difference between an AI stealing part of your work and a musician sampling part of another musicians song
1 Like
Simple, you need to ask permission to sample a song, and part of getting said permission is the agreement around any payment.
2 Likes
As @PHolder said, the musicians need to get permission and they have to pay royalties/a license fee, whether it is a cover version or the use of a sample in a new song.
Again, if the AI companies get permission and they pay the relevant fees, I don’t have any problem with that. If they are using illegal sharing sites to get the source material and using that without permission and without paying the licensing fee/royalties, that is not okay.
I’m late but I want to second and expand on this point.
We all look at things other people have made and incorporate them, intentionally or not, into our own internal model of creativity. Once a person sees an image, there’s no predicting what their brain will do with it.
I’ll never forget a certain educational experience: when I was a kid, a friend described to me a dream she’d had. It completely matched the plot of a book I had recently read. It wasn’t a popular book, so I was inclined to believe her when she said she didn’t even know it existed (and I still believe that’s probably the truth). We spent like ten minutes enthusing over what a weird coincidence it was. That was so very long ago, but it has stuck with me because it made me realize how flawed the concept of “originality” is.
1 Like