#Generative content

[prokopetz]

Producing generative content the way God intended: rolling on big stupid dice tables and deliberately misinterpreting the Rider-Waite tarot.

[savagechickens]

The Future of A.I.

[adastra-sf]

As generative AI spreads throughout publishing like a virus, Amazon is modifying how authors and other creators can use its Kindle self-publishing platform.

Amazon announced that it is "lowering the volume limits we have in place on new title creations" that users can add to the Kindle Direct Publishing platform daily to three, "in order to help protect against abuse," and it's possible that the number might be lowered again.

So now you can only write and publish three (3) Kindle books per day, folks.

When mega-corporations that otherwise exhibit no ethics begin to act against AI-generated content, you know things are getting bad.

Publisher's Weekly story: X

[leam1983]

Evolution

Remember how in 2005, people used to fuck around with a little chatbot called Cleverbot? Her responses were wooden, her tonal awareness was all over the map; she was easy to turn into an abusive shithead or to coax into agreeing with points of view that you couldn't realistically defend; and she notably couldn't handle a convo with another robot.

Now, we've got AI renditions of Werner Herzog and Slavoj Zizek going on endless tirades about the meaning of aesthetics and the underpinnings of their work in the empty void of a permanent exhibition space online, called The Infinite Conversation. While the semantics always break down if you really pay attention; things make sense on a sentence-per-sentence basis, and you see both AIs politely arguing and defending opposing viewpoints.

It used to be the best text-to-speech generator was what you heard on Emergency Alert System drills, and now we've got games like The Finals, that outsource the vocal underpinnings of their rapidly-changing game structure to AI, to avoid having to constantly re-hire the same voice actors to provide color commentary for game modes that might not exist six months down the line.

The kicker is their responses are eerily lifelike, with pauses, chuckles, scoffs, changes in emphasis that fit with events in a given match - and there's none of the choppiness you'd associate to text-to-speech à la Microsoft Sam.

I'm not condoning the practice, mind you; I'm just amazed that we've reached a point where, if you want, you can have a pair of RTX 4090 cards momentarily think like they're a peppy British-Korean lady and a snarky American man plastering fake cheer over live fire-enabled spectator sports set in the Distant Spacefuture.

[razehider]

this doesn't seem to be widespread knowledge around here yet but there's a big trend among dogshit content scraper accounts to grab a real photo (usually of ✨Aesthetic Nature™✨ or something similar, which is why it's relevant to me) somewhere, and recreate it using AI to avoid crediting the photographer. this can even trick people who are somewhat familiar with the subject matter if they're not paying attention but looks incredibly wrong upon closer inspection

here is some complete garbage as an example. because these "photos" are not completely made up by AI, people into spiders know the species and will recognize their features without looking closely, getting tricked in the process. if you know spider anatomy and look closely though, both of those look like utter abominations. the original photos these two were based on are here and here, by the way

these just so happen to be things i'm familiar with and i would probably get easily fooled by AI recreations of plants or fish or whatever. my point is that if you're not an expert on everything that exists you're not immune to these, so i would probably recommend caring about photo sources unless you actively want to look at this repulsive trash

[ai-super-babe]

Sexy AI-Generated Babe

[neuraldamsels]

Jess on fire

Source: deviantart.com

[thewarinourstarwars]

I want to weigh in on this after seeing this for myself.

I use em-dashes religiously and I once thought, well that’s unfair.

Then I had someone submit to me a piece of AI writing (and lie about it) and I knew immediately because folks—they didn’t use it correctly at all.

I realize we are on the autism website so I will clarify—when people are saying that em-dashes are the sign of AI, they mean non-sensical and incorrectly used em-dashes.

It is very very obvious when it’s in play.

Current writing advice I'm seeing on TikTok and Insta is telling authors to stop using em dashes in their work because, "AI uses em dashes so people will think you've used AI."

Y'know, the AI that was trained on the stolen work of real authors?

Anyway, I will not be doing that. What I will be doing, however, is adding a note at the start of all my books that no AI was used in the creation of my work because I, the author, did not go to university for four fucking years to study English literature and linguistics only to be told I can't use proper grammar because someone might think a robot wrote it.

Fucking, insane.

[rca-ryzies-ralley]

I think everyone should take note of the storytelling in the Ithaca premiere during luck runs out

[cozylittleartblog]

"content creator" is a corporate word.

we are artists.

[mckitterick]

fantastic assignment

but I'm concerned about how, even after this experiment, they're saying things like "revert to where AI is." this suggests they're still thinking of a statistics-and-database-driven algorithm as akin to the human mind

if we ever see true, Strong Artificial Intelligence - and I'm of the opinion we will in the next decade or two, once corporations stop wasting resources pretending the current nonsense they're producing is AI - its "thinking" will have little relationship to ours

[adastra-sf]

"The Wizard of AI"

We're just concluding our fall "Writing in (& about) the Age of Artificial Intelligence" workshop at Ad Astra headquarters (critique weekend is coming up!), and one of the participants shared this in our Discord. So naturally we have to share it with you, too, because relevant.

Artist and media critic Alan Warburton has released a brilliant new documentary about the evolution and state of artificial intelligence and its impact on creators, The Wizard of AI.

"I would say 99% of it was made using generative artificial intelligence tools," the 43-year-old filmmaker told The Guardian. Could his 20-minute film be the world’s first AI-generated documentary?

"I’m taking a leaf out of the AI hype playbook there," he admits with a laugh. "In truth, there is never going to be a first truly AI-generated documentary, because it always will involve labor of some kind. Labor is what makes it watchable."

The film does a fantastic job of sharing what he calls "wonder-panic" through its very existence.

One take-away after watching:

If AI has already "created" more images than humans have, shouldn't there be less to fear about their stealing human art? Surely they've polluted their own fishing waters, so to speak, so will AI "art" become increasingly inbred and thus (hopefully) irrelevant to the creative world?

We'll have to wait to see, because surely AI trainers at working hard at this problem, and big IP owners like Disney are, too. So perhaps it'll only be small artists who remain at risk of their work being stolen to train AI generators.

Warburton wrote the script, and says AI text generation lags far behind imagery. So at least there's that. For now.

Perhaps the most powerful part of the film is the closing "In Memoriam" section, listing all the creatives who didn't contribute to making it because AI took their jobs:

We strongly recommend everyone check this out, especially creatives, to get a glimpse at the state of AI in our daily lives:

"The Wizard of AI" on Vimeo

[hollis-art]

sparkle sounds

[goldensunset]

people who go through the main tags of big and tumultuous fandoms looking for new fresh good posts to reblog are essential to any circle. they’re like true hunter gatherers leaving the safety of settlement and braving the unknown wilderness to find food for the flock. they risk their lives every day and will come back with a few scratches at best and severe psychological damage at worst

[leam1983]

There's mitigation measures in place, unfortunately. Even attempts to "poison" datasets with Glaze aren't as effective as some people hope it'll be.

We're looking at an arms race of sorts, and OpenAI and its competitors are more than motivated enough to hop ahead of the development of mitigation tools like Glaze. You'll kill off tertiary generators running off of deprecated Stable Diffusion imageweights, sure - but you aren't likely to make a dent in Midjourney or DALL-E.

The irony is that you should be supporting the smaller guys running off of open-source tech that can actually be used to create legal imagesets, but that even these small fries are typically more interested in immediate profit than in working with the EFF and other bodies to finally bring about a decent means of accreditation.

I realized that with the way tumblr is set up it would be very easy to make a side account, opt into 3rd party shit, fill the queue and tag a post with every popular art tag I can think of and see what happens if I feed the robots some ducks real quicklike