The Weird World of AI Text-to-Image Art Generation

Unwary Traveller

Unwary Traveller

Acolyte
The ability to generate AI art by entering a text prompt (e.g. "a mouse driving a tractor") has been developing rapidly into something fairly astonishing, and I suspect anyone browsing the internet this year has been recommended a video or seen an article about it (there was some digital AI-assisted art posted in this thread - https://www.ligotti.net/showthread.php?t=13545). These are some interesting videos I've seen recently:

In case you missed all this hype in a haze of ennui, the first video is a general introduction focusing on DALL-E 2

How This A.I. Draws Anything You Describe [DALL-E 2] | Coldfusion
[ame]https://www.youtube.com/watch?v=U1cF9QCu1rQ[/ame]

Since the end of September, DALL-E 2 has become available for general use with no wait list. It's a very simple interface accessed through their website. When you sign up you get 50 free credits, meaning you can enter 50 text prompts, each of which will generate four images, which you can download and use how you want. After that you get 15 free credits per month, and you can buy more at a cost of 115 credits for $15. To sign up you need an email and phone number (verification codes will be sent to each; presumably the addition of a phone number is to discourage multiple accounts to get more free credits). It's very easy to get started and have some fun playing about using the free credits - try searching for "DALL E-2 prompts" to find videos or articles explaining how to get better results from your text inputs. Some words are restricted to avoid the generation of violent or sexual imagery. For other AI art generators see this article - How-To Geek: The Best AI Image Generators You Can Use Right Now.

The next video is on more of a weird theme, showing the potential for strangely disturbing AI imagery:

The Disturbing Art of A.I. | Nexpo
[ame]https://www.youtube.com/watch?v=i9InAbpM7mU[/ame]

Midjourney is another popular AI text-to-image generator. It runs within Discord, making it slightly more complicated to access if you are unfamiliar with that, but there are plenty of videos explaining the basics of how to set it up and use it. I've not tried it yet, I think you get 25 free credits, after which a very limited subscription (in terms of the number of images you can generate) is $10 per month and an unlimited subscription is $30. Not cheap, but you can see why people are fascinated by it. The next video shows a variety of Cthulhu images generated with the text prompt "horror artwork of H P Lovecraft."

100 AI generated images THE CALL OF CTHULHU tribute to H.P. Lovecraft | Midjourney AI
[ame]https://www.youtube.com/watch?v=DwyoIS8p6xI[/ame]

This video showcases a more varied selection of fantastical art, showing the detailed prompts used to generate them:

Ai Generated Art | MidJourney Top 100
[ame]https://www.youtube.com/watch?v=h4E8cw1JHlo[/ame]

Another AI project which has become available recently is OpenAI's ChatGPT, where you can chat with an AI on their website through text inputs and responses, in the way you've probably seen reported in the past. It's surprisingly sophisticated and the possibilities of this are wider than you might expect:

ChatGPT is Frighteningly Good | Enrico Tartarotti
[ame]https://www.youtube.com/watch?v=2lzQD-7lZJU[/ame]

This video is slightly mindblowing - using ChatGPT to generate descriptions which you can use as the prompts for the art:

ChatGPT + Midjourney V4 Showcase: Insane Combo! | All About AI
[ame]https://www.youtube.com/watch?v=hUyftBkfiOs[/ame]

Inspired by this, I asked ChatGPT to "think of some creative ways of using prompts to generate strange, weird, alien AI art pictures." This was its response:

Here are a few ideas for using prompts to generate strange and weird AI art pictures:

Use prompts that describe bizarre, otherworldly landscapes, such as "a planet covered in endless fields of glowing purple fungus" or "an ocean of acid where the sky is a boiling, stormy red".

Use prompts that describe strange, alien creatures, such as "a six-legged, tentacled beast with a bulbous, glowing head" or "a pack of predatory, humanoid insects with razor-sharp mandibles".

Use prompts that describe surreal, dream-like scenarios, such as "a city where the buildings are alive and constantly shifting" or "a forest filled with ghostly, translucent figures".

Use prompts that describe abstract, nonsensical concepts, such as "a color that cannot be seen by human eyes" or "a sound that causes the listener to hallucinate".

Use prompts that combine multiple strange and weird elements, such as "a cityscape where the buildings are made of living, pulsating meat" or "a landscape of floating, interconnected brain-like structures".
Click to expand...
It's starting to feel like I'm living in the future...

I'd be interested to hear peoples' experiences of this technology, either initial impressions (I tried DALL-E 2 yesterday and this is all very new to me) or from more experienced users. My ability to paint or draw is extremely limited, so I find this ability to generate pictures from text very inspiring. Try it!
 
It is very, very spooky. I wonder where it will lead. I sense madness. And an
approaching social nightmare.

Although extremely impressive, I don't think it has the natural beauty, and healthy balanced profundity of the great human artists.
 
The images in the video: Ai Generated Art | MidJourney Top 100 are as good as any art that I have ever seen.

The implications of this technology are difficult to fathom.

Music and literature will be next.
 
Last edited:
All the best AI art I've seen is heavily based on something that was already great, so in a way it resembles a patchwork of good songs mixed together in a hopefully coherent way.

But almost all of it has a very plastic texture if you look closely. The novelty wears off quickly and I think you actually have to work at it a bit to get anything good from it. Of course, it might advance and get rid of these limitations but right now I'm not worried at all other than the copyright aspects people have been bringing up and seeing so much of the bad stuff everywhere is wearying. But I seen someone compare this to the panic about photoshop (and perhaps the way people panicked over synthesizers and photography?)
 
And consider that CGI in films still rarely look convincing or good. It takes a lot of work and perhaps more importantly: TIME that big studios won't wait, so they look awful.
 
There is a lot of fuss about it but I don't see it as a threat at all:

1) Art as commodity: there will be a hype around buying AI-generated art but that will sizzle out because the abundance of AI art makes it less appealing as wealth signal. Therefore, AI generated art will push the value of human-generated art higher.

2) Art as communication: what does AI communicate? Nothing. It has no context of feelings, culture and beauty. Looking at AI generated art is cool but it gets boring after a while. What else can we discuss about its art other than its AI-generated?

3) Art as idea: in a way AI furthers the dead-end path artists have been traveling which is to separate ideas and craftsmanship. You have artists who sell ideas and leave the art-making toward others. Art which is purely shocking and pretentious like a used tampon on display. AI generated art is just more of this.
 
মোস্টবেটসে বিশেষ বোনাস এবং প্রচারমূলক অফারগুলি পাওয়া যায়, যা আপনার জুয়া খেলার অভিজ্ঞতাকে আরও আকর্ষণীয় করে তুলবে। বিস্তারিত জানতে এবং বোনাসের সুযোগগুলি দেখতে এখানে https://mostbets-bd.com/bonus/ ক্লিক করুন। আপনার জয়ের সম্ভাবনা বাড়ান!
 
Last edited:
It will, among other things, invoke the age-old question of "what is art?" I find it interesting in a somewhat general sense, that people are becoming more machine-like, and machines are becoming more human-like. Not in a literal sense, of course, but in some ways it is undeniably true. In some ways, if not many, people interface far more with machines than they do with actual human beings, which should or could induces terror in some people. It isn't a difficult argument to make the we are living in Hyper-Reality, as Baudrillard might have insisted.



We are living in a Trans-humanist's dream-like state where the lines between "reality" and fantasy, synthetic and organic, are being blurred, and I do know that people's reality testing, from a classical point of view, is being impaired to one degree or another by interfacing with endless streams of simulacra so consistently. I see it expressed in in countless ways in the people that I work with on a daily basis. For example, I can ask people questions and predict their answers with a remarkable degree of accuracy, primarily because before I ask, I know where they are getting their information from.



It's a long story for another time, but it is very startling what I see out there as a Therapist who works in private practice and because of the kind of work I do with the people that come to me to engage in the process.
 
Ascrobius said:
It will, among other things, invoke the age-old question of "what is art?" I find it interesting in a somewhat general sense, that people are becoming more machine-like, and machines are becoming more human-like. Not in a literal sense, of course, but in some ways it is undeniably true. In some ways, if not many, people interface far more with machines than they do with actual human beings, which should or could induces terror in some people. It isn't a difficult argument to make the we are living in Hyper-Reality, as Baudrillard might have insisted.



We are living in a Trans-humanist's dream-like state where the lines between "reality" and fantasy, synthetic and organic, are being blurred, and I do know that people's reality testing, from a classical point of view, is being impaired to one degree or another by interfacing with endless streams of simulacra so consistently. I see it expressed in in countless ways in the people that I work with on a daily basis. For example, I can ask people questions and predict their answers with a remarkable degree of accuracy, primarily because before I ask, I know where they are getting their information from.



It's a long story for another time, but it is very startling what I see out there as a Therapist who works in private practice and because of the kind of work I do with the people that come to me to engage in the process.
Click to expand...

I have been circling this subject for quite a while. As soon as I brush up on my Baudrillard (there are numerous YouTube videos on the subject) I look forward to reading this book, which I already own and is on my short list.

1839980710.jpg
 
I think that the pictures made by artificial intelligence over time will be more and more qualitative. But this process could take a long time. In the meantime, I want to recommend to you an interesting article. The website Parimatch is a resource that provides information about the top 10 football teams in Africa. It likely features detailed information about the teams, their achievements, players, and other interesting facts related to football on the African continent.
 
Last edited:
Nick Cave calls ChatGPT and AI songwriting ‘a grotesque mockery of what it is to be human’

“Songs arise out of suffering, by which I mean they are predicated upon the complex, internal human struggle of creation and, well, as far as I know, algorithms don’t feel,” he added.
“Data doesn’t suffer. ChatGPT has no inner being, it has been nowhere, it has endured nothing, it has not had the audacity to reach beyond its limitations, and hence it doesn’t have the capacity for a shared transcendent experience, as it has no limitations from which to transcend.”

I don't understand this attitude of self-importance.
 
I think we are going to see increasingly more people getting disturbed and frightened by their computers as we descend further into the uncanny valley. It should be a good time.
 
This should end debate as to influence vs plagiarism.

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
 
Last edited:
Really interesting discussion — AI text-to-image tools have evolved insanely fast, and it’s fascinating to see how accessible they’ve become even for people without a technical background. What stands out to me is how much the user experience matters now: simple prompts, clear interfaces, and fast results are what really drive adoption, not just raw technology.
That same focus on usability is something I’ve noticed on platforms like krikya app, where everything is structured so users can jump in quickly without feeling overwhelmed. Whether it’s creative tools or entertainment platforms, clarity and ease of use seem to be the common thread behind what actually works.
 
You must log in or register to reply here.
Back
Top