Jump to content

Recommended Posts

Posted

Splitting off, because I absolutely agree with the "Thank you Chaosium" sentiment and don't want to detract from it.

On 12/17/2022 at 4:06 AM, g33k said:

In case nobody has seen this one...

Art by AI ... Text by AI ...

https://time.com/6240569/ai-childrens-book-alice-and-sparkle-artists-unhappy/

aaaiiiieeeee !!!

Fascinating. As a technologist, AI art is amazing, and I love playing with it, but I absolutely see the concerns.

The story in Time magazine is clearly a demonstration of what is possible, and as such I think the "artists are not happy about it" is absolutely the point of the demonstration. Like "deep fakes", there are clearly ethical issues about making fake political speeches or fake racism that look like a real person saying terrible things to traduce them, but it's not the technology that is at fault. My mum said "why have they created something so problematic", but that's beside the point, like nuclear bombs, someone was going to do it eventually. I'm sure there are ethical ways to use AI to generate or manipulate art, but demonstrations like the Time story raise the issues in a tangible way that people can relate to. "It's here, we need to deal with it". I expect that that is what Reshi was doing.

Would an AI trained entirely on public domain, "old masters" art, Shakespeare, Austen, and Dickens, be ethical? Even then I can still see organisations choosing to take Chaosium's line on it, because we value the creators that supply us with great content, we are those creators. And the AI output would all be old-fashioned anyway.

  • Like 2
Posted

We have reached the threshold where we cannot believe what we see, hear and read. And if someone thinks that this isn't being used to manipulate the populace into buying more this or that, voting for x or y, or psychologically preparing for war, they are sadly mistaken.

It used to be that it was a fairly easy thing to tell the difference between fact and fiction and to identify the hype and hyperbole of propaganda. Now it requires a skilled technician to validate an image, sound or printed page. Even the slang terms for this kind of stuff is misleading... 'gaslighting', 'deep fake', etc.

I'm not a Luddite by any stretch of the imagination, though several failed products have led me to not be an 'early adopter' of most technologies. But at what point will we finally decide to value the human-created over the machine?

  • Like 1
Posted (edited)

I'm largely in agreement with this sentiment, but take exceptions  to some details ...

2 hours ago, svensson said:

We have reached the threshold where we cannot believe what we see, hear and read. And if someone thinks that this isn't being used to manipulate the populace into buying more this or that, voting for x or y, or psychologically preparing for war, they are sadly mistaken.

The entirely human-produced lies -- inherently natural-language, with a firm grasp on how to manipulate the emotion and intentionally engage with pre-existing biases; or with carefully-curated source-images, composited & re-contexted in specifically-calculated ways -- are tremendously more effective than this brand-new tech, with (most of the time) a substantive "uncanny valley" effect from many/most iterations of the AI output.  I don't think there's (very much) AI-created propaganda in use, yet; neither commercial ad's (although q.v. the model Shudu Gram featured by Balmain & Louis Vuitton, among others) nor political hackery ("Bermuda" is a "social media influencer" who used to be pro-Trump but now... isn't, I guess?  I don't follow "her" so I'm unclear on the details).

The time is coming (very very quickly, I think... possibly a few months? maybe a few years?) where this won't be true; the AI output will be substantively indistinguishable from humans' output; it's already very, very good!
 

2 hours ago, svensson said:

... It used to be that it was a fairly easy thing to tell the difference between fact and fiction and to identify the hype and hyperbole of propaganda ...

The thing is, it may have theoretically been "easy" ... but hundreds of millions of people haven't been putting in that minimal effort to do so.

Even before this latest round of (AI-driven) fakes, people would blithely repost fake meme's, disinformation, calls to violence &c ... when it agreed with their biases, or even "for the lulz" & to troll the "other side."

And as the oft-quoted saying goes, "“a lie can travel halfway around the world before the truth puts on its shoes” (n.b. it's not true that that is a Twain quote:  the original speaker is actually unknown.  Best explication I have found is here: https://professorbuzzkill.com/twain-lie-travels/ )
 

2 hours ago, svensson said:

...

I'm not a Luddite ... But at what point will we finally decide to value the human-created over the machine?

I trust you see the irony in your own remarks!

Once again:  I'm largely in agreement with the sentiment, but the specific formulation... well... "I do not think it means what you think it means."  Or you'd be driving a hand-crafted automobile, you'd have written your post on a hand-crafted tech device, etc.  Take a good hard look at your own budget -- what % of your $$$ are you confident have gone to the crafters for their hand-work, vs. to the output of automation?

I will argue that -- almost certainly! -- you yourself are "valuing" (spending more of your money) on the output of automation.

(excepting only the possibility of a mortgage (as homes are almost entirely handwork, still); but unless you are the 1st owner, the hand-crafters didn't benefit from your purchase, so I don't think that "counts" on the handcraft-vs-automation debate (instead, you're benefiting a 3rd axis:  capital/finance (unless you can buy a home outright) ).
 

Edited by g33k
  • Like 1

C'es ne pas un .sig

Posted (edited)
5 hours ago, PhilHibbs said:

Splitting off, because I absolutely agree with the "Thank you Chaosium" sentiment and don't want to detract from it.

Fascinating. As a technologist, AI art is amazing, and I love playing with it, but I absolutely see the concerns.

The story in Time magazine is clearly a demonstration of what is possible, and as such I think the "artists are not happy about it" is absolutely the point of the demonstration. Like "deep fakes", there are clearly ethical issues about making fake political speeches or fake racism that look like a real person saying terrible things to traduce them, but it's not the technology that is at fault. My mum said "why have they created something so problematic", but that's beside the point, like nuclear bombs, someone was going to do it eventually. I'm sure there are ethical ways to use AI to generate or manipulate art, but demonstrations like the Time story raise the issues in a tangible way that people can relate to. "It's here, we need to deal with it". I expect that that is what Reshi was doing ...

I too am a technologist, and also fascinated by the new tech, the new potential; and worried about the new risk.

I agree that "someone was going to do it anyway" and "it's here, we need to deal with it."
[  I don't think suppressing an entire technology is likely to be a winning strategy!  (q.v. EV growth -- e.g. 2022 EV sales for USA/Canada grew 49% year-on-year, while the overall "light vehicle" market shrank by 17% for the same time-period; and you KNOW that Big Oil & Big Auto would have quashed the EV market if they could have!  And see this interesting article & associated chart: https://www.visualcapitalist.com/visualizing-10-years-of-global-ev-sales-by-country/ )  ]

But I'm not sure I see (at this point) the general methods to "deal with it."  It looks very-much like a Pandora's Box, and we just have to settle in for the bumpy ride, knowing there WILL be some problems; we may have some patches and spot-fixes, but no real "solutions."

Honestly, though, I think you give far too much credit to Reshi -- iirc he never said or implied he was trying to highlight any ethical issues; he just wanted to explore the exciting new tech, and make something cool for his friends' kid; which, y'know... fair enough!  If I had a really-good friend with a new kid, I might have a similar impulse myself!
 

5 hours ago, PhilHibbs said:

...

Would an AI trained entirely on public domain, "old masters" art, Shakespeare, Austen, and Dickens, be ethical? Even then I can still see organisations choosing to take Chaosium's line on it, because we value the creators that supply us with great content, we are those creators. And the AI output would all be old-fashioned anyway.

Gotta point out that a great deal of that "old fashioned" content is absolutely desirable to many gaming audiences.  Austin, you say?  Chaosium has just published Regency Cthulhu, I point out!  And I think we all realize that the art-direction for Runequest leans heavily upon ancient art & artifacts!

One aspect of the anti-AI argument (an argument I am broadly sympathetic to) is that the AI is "scraping" the content of living artists, without their consent or compensation.  I somewhat regretfully conclude this isn't necessarily a sound criticism:  things like "midjourney" *DO* scrape from living artists, but (unless you key-word to particular artists, well-known titles/works, or styles) they scrape very-broadly from MANY artists, and (that I have seen) synthesize something rather different and NOT really "stealing from" anyone in particular.  What, really, is the difference between "scraping" many artists vs "studying other artists for inspiration & technical learning"?  Is the AI not doing -- broadly speaking -- the same thing that living artists do?   This is, BTW, a very-genuine question, not rhetorical:  I'm not a trained artist, nor an AI-programmer (let alone both!), so I don't know.
 

Edited by g33k
an addendum, for those interested who hadn't seen -- https://www.cnn.com/2022/09/03/tech/ai-art-fair-winner-controversy/index.html

C'es ne pas un .sig

Posted
On 12/19/2022 at 4:30 AM, PhilHibbs said:

Splitting off, because I absolutely agree with the "Thank you Chaosium" sentiment and don't want to detract from it.

On 12/16/2022 at 9:06 PM, g33k said:

Thanks Phil, good topic.

... remember, with a TARDIS, one is never late for breakfast!

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...