TSR The Full & Glorious History of NuTSR

Steampunkette · May 25, 2023

Sacrosanct said:
This article seems relevant. A SC ruling on Andy Warhol using someone else's art to create his own from (which is what AI does). Ruling? You can't do that.

https://www.npr.org/2023/05/18/1176881182/supreme-court-sides-against-andy-warhol-foundation-in-copyright-infringement-cas?fbclid=IwAR1RS-1fGQybUWtffWdTWB3OrKMicLeeZxNWma5ht274tdY4EJTV3uxvRPg

The issue with the Warhol thing is that it uses someone else's art -directly-. It isn't influenced or guided by someone's art. He just re-printed it in different colors.

The AIs that are on debate don't sample someone else's art and composite it together. They learn by looking at a bunch of art and develops their own based on what it's learned based on a preponderance of the data.

If 200 artists do a portrait as face-on with symmetry and 10 do the portrait as looking off to the left or right slightly it learns people prefer portraits face-on and creates more that look directly at the 'camera'.

The underlying assumption that AI art is just stealing people's artwork and compositing it together like a puzzle is the reason the current challenge against it is likely to fail. The artists bringing the legal challenge against it didn't understand the technology when they brought the case, and expert testimony is going to eviscerate their case as they've built it.

A more reasonable argument has been brought forward, that the artists didn't agree to let the AI get trained on their artwork, but that brings with it an entire -separate- legal issue: What artist specifically allows their art to train any other artist? The number of art classes based on teaching the styles of DaVinci, Van Gogh, and even Rothko that exist teach and train based on people's pre-existing art. Entire classes are built around mimicing specific -paintings- to convey the style.

Heck, the number of young artists who learn to draw by pausing Anime to copy the style of Akira Torayama or other artists is enormous. Would they suddenly also have a case to sue the kids and teens? Or does AI get a special carve-out because of the intention of a programmer compared to the AI lacking intention. And what are the repercussions of teaching any AI anything if that carveout can be expanded to cover other topics.

There's a lot more to it than appears on the surface.

EDIT

Just went back to copyright alliance for the first time since early March:

The Class Action Lawsuit -did- update their arguments to include the "We didn't agree" option. But they still hold to the "They composited our art and stole our profits" line. The former could reasonably result in punitive damages. The latter is gonna get torn apart barring an incredibly sympathetic judge who refuses expert testimony... in which case compensatory -and- punitive damages could result.

But I doubt it.

Umbran · May 25, 2023

Steampunkette said:
A more reasonable argument has been brought forward, that the artists didn't agree to let the AI get trained on their artwork, but that brings with it an entire -separate- legal issue: What artist specifically allows their art to train any other artist?

Objection, your honor! This presumes the conclusion that the machine qualifies as an "artist", which has not be established in law or this courtroom.

LordEntrails · May 25, 2023

Isn't there a whole other thread to discuss AI art? I'm thinking this is not the right thread for that discussion. Please.

Steampunkette · May 25, 2023

Umbran said:
Objection, your honor! This presumes the conclusion that the machine qualifies as an "artist", which has not be established in law or this courtroom.

That could lead us down the rabbithole of defining personhood (An artist is a person who creates art) -or- trying to define art (The machine produces something that is art-like but not truly art)

Which one do you wanna dictate for future generations?

After all, what's the real difference between teaching a robot by showing it a thousand pictures and doing the same thing with a human?

Umbran · May 25, 2023

Steampunkette said:
That could lead us down the rabbithole of defining personhood (An artist is a person who creates art) -or- trying to define art (The machine produces something that is art-like but not truly art)

If the judge allows the argument, he's tacitly doing so anyway, so...

Dire Bare · May 26, 2023

Umbran said:
Objection, your honor! This presumes the conclusion that the machine qualifies as an "artist", which has not be established in law or this courtroom.

Not giving machine intelligences full consideration as personhoods . . . . . is what leads to machine uprisings and the destruction of humanity.

You'd think nerds would be more aware of this danger . . . .

Umbran · May 26, 2023

Dire Bare said:
Not giving machine intelligences full consideration as personhoods . . . . . is what leads to machine uprisings and the destruction of humanity.

You'd think nerds would be more aware of this danger . . . .

As if teaming up with a megacorp is evidence of their benevolence?

You'd think nerds would be more aware of this danger...

Wincenworks · May 26, 2023

Steampunkette said:
That could lead us down the rabbithole of defining personhood (An artist is a person who creates art) -or- trying to define art (The machine produces something that is art-like but not truly art)

That was already settled in the US thanks to PETA trying to use copyright to backdoor in giving human rights to animals.

Only a human may create art. The AI is not a human, it is a tool.

That AI art generators operate on prompts means that the only thing that could foreseeably be copyrighted is the prompt, which is almost certainly never going to be unique and complex enough to attract copyright.

All of the artistic decisions in AI art are made algorithmically, therefore are matters of fact and data, and cannot attract copyright. Only human decisions can. There is no foreseeable future where AI attracts copyright on its own, only the changes made by the humans do.

So in this case, what OSR Games LLC has "copyrighted" is the placement of the text overlay on the top - which is kind of hilarious since it implies that he thinks someone else wants to make a sign for his business.

Steampunkette said:
After all, what's the real difference between teaching a robot by showing it a thousand pictures and doing the same thing with a human?

Well, one, that it's a human who operates on a lifetime of information, free will, etc. The computer doesn't do any of that, it doesn't understand anything, especially the concepts of the art that it is. It doesn't know what a person is, what ramen is, what eating is - it just scans information that is loaded into it and interpreted for it by humans and then generates something that it thinks matches the patterns.

This is why AI art continually generates highly polished gibberish, instead of varying degrees of technical understanding work that is based around a concept - something with AI "artists" are struggling with now that they've been trying for long enough to get it to create a specific image and are finding the AI can't approach it because it doesn't understand the concepts, and can't enter into a conversation with you.

This weird "it's like a person" rhetoric only makes sense if you want to also assign pseudo-human status to:

autocorrect
spell check
this forum software
Rhoombas
spreadsheets
calculators
light sensitive switches
rice cookers

The delusion that something which does something we associate with a human (even something as looking like it has eyes) is alive/human etc is an ancient one based in our attempts to build narratives that make sense of the patterns we observe, but the AI art generator is no more an artist than a pet rock is a listener.

This particular delusion compounds because people assume that if an image is created, and its complicated - it must attract copyright. This is the general rule, but there are many exceptions such as if you recreate something that is already in public domain, do it under circumstances were you agree not to apply a copyright (part of an open license) etc. AI art has created a whole new branch of this because they've made an algorithm which interprets data to make an image.

Thus is can violate copyright, but it can't create images that attract copyright and more than the contents of the periodic table can.

Not unlike how it seems Justin thought he could bypass the copyrights on the images that make up TSR trademarks by drawing copies of TSR The Game Wizards logo and the Mike Bell Lizard Man - deliberately trying to recreate the images as faithfully as possible and then later hiring someone who can draw (reminder, Justin is supposed to be a tattoo artist) to make the current one which is a lovely homage work by Diesel, but is definitely not the image Justin submitted to the Trademark office.

Steampunkette · May 26, 2023

Fair point about the ape. Though when the Centauri come to Earth and can't hold copyrights on their Faqooli music I think they're going to be super upset and possibly conquer us.

As far as "Free Will" and "Lifetimes" are concerned, though, that's some dubious ground to stand on, philosophically. Much less legally.

The only point I actually take issue to in your post, however, is the use of the word "Delusion". Because it ignores a basic tenet of language and reality in favor of trying to force things into a narrow conception of reality. Specifically: Everything is made up and the points don't matter.

Personhood is often ascribed to things that are no more -human- than other things. Like an AI and a Rock in your example. However personhood can -only- be ascribed to things regardless of their actual function toward personhood because personhood doesn't mean "Human". We -personify- objects and animals. We ascribe to them humanlike traits. We talk about temperamental machines for example but no one thinks their car is actually human even if it has anthropomorphic traits ascribed to it. No matter how gently we talk to it while turning the key and begging the engine to turn over.

However. In the case of Artificial Intelligence those humanlike traits, the personification, is the -goal-. It exists exclusively as an attempt to create a thinking device, something that does what humans do in the way that humans conceive it.

And while Stabby the Roomba certainly isn't Data from Star Trek, that is the eventual goal. Ascribing personification to AI is markedly different to personifying a rock.

After all, a rock can't gaslight you about your marriage like ChatGPT can. Does it understand what it's doing? Not yet.

That yet should either excite or terrify, depending on whether you enjoyed "The Measure of a Man" from NexGen.

Wincenworks · May 26, 2023

Steampunkette said:
Though when the Centauri come to Earth and can't hold copyrights on their Faqooli music I think they're going to be super upset and possibly conquer us.

Yup, things sure will be different when they are different - fortunately for us the law is constantly evolving and operates on rules, interpretations and facts - not sitcom logic.

Steampunkette said:
As far as "Free Will" and "Lifetimes" are concerned, though, that's some dubious ground to stand on, philosophically. Much less legally.

No, it's very clear legally. A human makes choices and is responsible for those choices, but a human can also point to any events in their lifetime as part of a cause for their choices. A machine is built and operates on set rules given to it by a human, directly or indirectly. An animal makes choices, but does not operate in human frameworks so does not have human rights and responsibilities.

Personally I think art made by rats is way more fulfilling than most of the crap that Andy Warhol farted out to make a quick buck, but I'm not going to pretend the rat can meet the standards required to attract copyright in any region.

Speculating on if an alien life form would is pointless until we meet the life form and see how it reacts to us, or even how we try to communicate with it.

Steampunkette said:
After all, a rock can't gaslight you about your marriage like ChatGPT can.

Sure it can, you can decide to have a conversation with it - react to its imaginary judgments and come to the conclusion your wife is cheating on you. History is literally full of people making decisions because they believed inanimate objects could give them secret insights or guide them - omen readings, dowsing rods, rune readings, palm readings, pyromancy, tyromancy, and all manner of weird seances. People also regularly take fundamentally irrelevant data and try to claim it's massively significant to their situation - see Nostradamus.

Doing it through a computer isn't even new, you could get questionable advice on any personal topic of it thirty years ago if you owned a Soundblaster card.

Steampunkette said:
Does it understand what it's doing? Not yet.

It's never going to because the research isn't going in that direction and nobody has a good model for even trying to make it go into that model - all it does is pattern match data and try to produce a response based on that pattern - just like a kid trying to trick fireflies by flashing a light at them, or a bird watcher doing birdcalls, or the LCD display on your phone when the camera is operating.

An algorithm designed to match and respond to patterns can never do more than match and respond to patterns, not matter how much your imagination tells you that it's learning to be smart or becoming alive.

It's important to understand this because it's helps prevent against charlatanism such a supporting pivoting to AI creative (a scam to undermine what actual creative people get paid - thus contributing to the WGA strike), new AI regulation laws (which are to lock out competition by making them prove they're not creating the type of AI we can't make exist before they can start work), supporting Robot Lawyers and other substitutes for high stakes professionals, etc.

This isn't unique to AI or even special, lots of people believed in the potential of the technology that Theranos was claiming to be developing to - not because it was workable, or they could show how it was going to work, etc, but because the idea of us making this technological leap to where we can do an auto-diagnosis with nothing more than pinprick was a seductive idea that people (including Elizabeth Holmes) wanted to believe in.

And this context, it's important to understand it because its worth understanding the absurdities of a lot of nuTSR's fiascos.

Justin claims, and he and his (tiny) fanbase seems to genuinely believe that he's "making games" because he's published a book that copy-pastes a lot of stuff from various old editions of D&D, includes terms like ThAC10, tables of numbers, spells and names of monsters etc. He feels that because it ticks off these boxes - it must be a game and good. He's wrong not because games don't have these kinds of content, but because he doesn't understand why they do and doesn't appreciate how those elements serve the greater purpose.

They're an example of the kind of content you get when you mistake pattern matching for creativity.

Steampunkette said:
That yet should either excite or terrify, depending on whether you enjoyed "The Measure of a Man" from NexGen.

I'm more of a Dune guy, but I don't see any evidence that the Butlerian Jihad is upon us - as much as I like to threaten my housemate's Rhoomba with it.

TSR The Full & Glorious History of NuTSR

Steampunkette

Rules Tinkerer and Freelance Writer

Umbran

Mod Squad

LordEntrails

Hero

Steampunkette

Rules Tinkerer and Freelance Writer

Umbran

Mod Squad

Dire Bare

Legend

Umbran

Mod Squad

Wincenworks

Hero

Steampunkette

Rules Tinkerer and Freelance Writer

Wincenworks

Hero

Similar Threads

Voidrunner's Codex

Related Articles

Voidrunner's Codex