After leaving Bing imagine creator running for hours (overnight) I came back to find more Koalas in knitted hats! It gave me four koalas and I chose one to enhance with their program.
Look it’s the same couch! and half of the Koala is gone…
This baby koala has different claws then the others. My 7 year old tells me these are all baby Koalas because their ears are spiky.
I decided to play with Koala 3 with what Bing had to offer.
I played with the lighting and gave it a boarder. Bing was really slow and there wasn’t much I could do without paying for more. It did have a lot of templates to turn the imagine into a card or announcement. My suggestion is to give it 12+ hours to generate the imagine.
The first image generator that I decided to use was Stable Diffusion.
The following statement was what I asked for it to provide me with: create an image of someone cooking Pakistani food in a kitchen
This was the image I got back:
Personally, this wasn’t enough for me (not in terms of dishes, I mean the vibe all together).
So I gave it the following description: create an image of someone cooking a variety of luscious Pakistani food in a kitchen decorated with fancy plates and color around the room
The image I received back was
I tried about three more times, and I was disappointed with the results. It kept generating the same type of dishes and the architecture in the room was the same as well.
The next image generator I tried was Craiyon.
I fed it the following description: create me an image of a someone cooking a variety of mouth watering Pakistani cuisine with a fancy and colorful appearance overall
The image I received from them was the following:
I found this website to be the most inaccurate and frustrating. It displayed me 9 images each time, and none of them met the criteria.
This is why I always say that you can never fully rely on AI.
AI image generators is a threat to artists. The way I see it as is that artists put thought, effort, time, blood, sweat, tears and so much work into their pieces. Whereas for these image generators all someone is doing is sitting behind a screen and coming up with a description to feed AI. I find that in this way, the concept of art if gone… lost in translation.
AI art has always been a strange topic to talk about. I am practicing drawing, which is extremely difficult and seeing AI create art in seconds is upsetting. But I still believe that AI can never recreate the personal aspect that people add to their creative work. This assignment was fun, but it made me realize how important language is to the AI, as it cannot create the image you have in your mind exactly the way you wish. I used Stable Diffusion first and entered the prompt, “Blue butterfly in a suit.” I thought the outcome would be a butterfly with a human body, wearing a suit. But to my surprise, it was something else.
This was ridiculous. I tried using Craiyon, but every outcome was a regular butterfly.
I tried adding more detail in the text and it got very strange. My description changed to, “blue butterfly with a humanistic body wearing a suit and holding a cigar”. This was the result on Stable Diffusion.
Pretty disturbing stuff. Though it is not what I asked for, it still has a unique aspect to it.
Craiyon was unable to give a correct result. It stuck with a basic image of a blue butterfly. It is probably the worst result out of everything else. While the Stable Diffusion result was completely different, it was still creative enough to keep me interested. This is the Craiyon result.
Overall, it was fun to try new descriptions but the process of creating these images are questionable. The act of taking other artists work without permission should be pressed on. I think that use of AI art can be interesting as long as it is purely for creative purposes. I think using AI for profit is ethically wrong, since artists have not given permission for these algorithms to use. I think the use of AI is inevitable, but people need to set their limits on how far we should allow this to act.
I started this assignment by choosing Stable Diffusion and Bing as my AI photo generators. My initial prompt was “Grateful Dead”. I got the following images:
I guess stable diffusion is not as aware of the Grateful Dead!
I advanced the prompt to “The Grateful Dead playing in front of a large crowd”.
These are the images I got:
Stable diffusion getting there, but still not too sure.
My next prompt was “The Grateful Dead in the style of Starry Night”. It was rejected by Bing so I changed it to “The Grateful Dead in the style of Picasso”. You know the drill – here are the images:
My “worst” image is the first stable diffusion image since it did not include anything related the Grateful Dead. The most impressive image has to be the original Bing Grateful Dead image since it had to pick up on the dancing bears and the psychedelic Dead theme.
I imagine AI generated art can be a threat to artists in some aspects, especially concert posters. After doing this assignment and seeing some similarities between the posters from concerts and the image that the AI generators put out, I definitely believe that could threaten some artists.
This assignment was definitely a new one for me. I’ve mostly tried to avoid AI art generators as I prefer making my own artwork but also because of all the negative things you hear about it in the news. So it took a bit more trial and error than it may have for others for me to get each idea I had to work.
The worst image I would say I got was when I prompted: cupcake on a plate with a yellow car fire with explosions behind it in the style of Salvador Dalí. I got images that were mostly just whimsical but not in Dalí’s style which makes me think that the software may have a sort of idea of who he is and what his work looks like but not enough to recreate his style properly. This reminded me that these A.I art generators can only create on the base of what they were trained with which limits it so much in what it produces. While my prompts changed every time, the same art style and usage of the cupcake and fire remained almost the same.
I would say the most impressive one generated would be when I asked for: whimsical cupcake on a plate with a yellow car fire with explosions behind it. Specifically the second stable diffusion one with the two cars behind the cupcake. I think this one works well because I put the style of the art in front of the rest of the prompt and was more specific than I had been in the others. I like how it captured whimsical as well with the odd car on the right with the two roofs and the art style that the fire on top of the cupcake is like a candle.
After this exercise along with the reading before discussing the usage of A.I. art today, I would definitely perceive these generators as a huge threat to artists. While it may not react to a prompt or idea with the same creativity a person might, it is still capable of producing unique images when the prompts are specific and adapted to its language. If someone’s art is uploaded then their entire style is stolen essentially and anyone with access to the generator can make things faster and of better quality often than the artists themselves which puts not only the jobs but also the motivation of these artists at risk.
Step 2: Build your base
Step 3: Add some flair
Step 4: Transform it
Program
Cupcake on a plate
Cupcake on a plate with explosions behind it
Cupcake on a plate with a car fire behind it
Cupcake on a plate with a yellow car fire with explosions behind it in the style of Salvador Dalí
Cupcake on a plate with a yellow car fire with explosions behind it in abstract style
Abstract cupcake on a plate with a yellow car fire with explosions behind it
Whimsical cupcake on a plate with a yellow car fire with explosions behind it
This first image of drake was created on stable diffusion. This image is clean and simple with a realistic look.
This image of drake was created on craiyon. This image gives more of a cartoon look and not so much a realistic look compared to the first generated image of drake.
I felt like this was the worse image that was generated. This image came from craiyon. I asked craiyon to generate a photo of drake holding a puppy and this is what I got. The reason why I think this is the worse image is because one Drake facial expression is unappealing, he looks disturb in the photo. The second reason is the puppy in the photo does not look realistic at all and on top of that it almost looks like a robotic puppy.
I felt like this was the best generated image of drake. This image came from stable diffusion. I asked stable diffusion to generate a photo of drake holding a puppy and this is what I got. I think this is the best generated image because it gives off good vibes. Drake looks happy and his smile gives off an overall happy mood. The puppies also add to the good vibes as they are adorable and joyful.
I think AI can be a potential threat to artist. For starters an AI can almost create anything you ask it to. Whereas an artist is limited to what he or she can create. AIs can modify to enhance an image you ask them to, and this puts artists at a disadvantage because an artist may be incapable of doing what the Ai can.
I really enjoyed this assignment. I tried a few different prompts. Spider-man from the Marvel comics did not work. Trust me or don’t and see what an Adobe Firefly gives you. I settled on something less startling Koala on a blue couch wearing a knitted green and purple hat. I tried to use a few other AI image generators but I either had to pay or I am sill currently waiting on them to finish generating. At the time of typing this it has been at least an hour. So, I will show you a selection from Adobe Firefly.
I really like this one because it is like the koala is posing for a portrait. The couch is fancy and is outside with eucalyptus trees in the background. The recuring problem is shown here. Half of the koala is missing. With the scale of the couch and a koala half of him is gone or inside the couch.
This one I do not like. I asked for the style of cut paper and knitted. I forgot to remove the knitted style from the prompt. It gave me a weird blend of flat paper and stuffed animal. The koala is clearly being eaten by the couch. I don’t understand the need to cut paper in its ears. the couch ends and begins on the right side of the imagine. Look at just the lines of the couch and you can see it doesn’t make sense.
I love the style of this imagine. The style I asked for a painting, acrylic, watercolor, geometric pen, and photo. I was trying to get something that did not work at all. It gave me a fun blend and the koala some pens. But sadly, once again half of the little guy has been cut off. The couch also has the weird begins and ends thing happening on the right of the imagine.
I was surprised that by adding more style prompts I got a more pleasing imagine than with less style prompts. I thought it would have been the other way around.
I started this activity with hesitation, really not excited to be creating AI art, so I must preface this by saying that the only reason I am doing this is for a grade, but also the willingess to learn, but that I really don’t see anything good about using AI to create art.
I wanted to work with my cat and create images of her and see what it could come up with. I used Stable Diffusion for this.
The first prompt I input was “calico cat” because, well, my cat is a calico cat. Here is what it spat out:
It doesn’t look like my cat, and it definitely feels more like a tabby calico than just a calico, but I knew I’d have to refine details to get closer to the image I wanted. Already, I was feeling a little weird looking at it. It has that flair that AI art has that makes me uneasy.
Next, I input “closeup of a shorthair calico cat with green eyes with a bigger build” and this is what it gave me:
It’s closer. The cat definitely looks a little more thin than my beautiful, chunky (said with love) cat. The patterns are still a little off, but I didn’t give it specifics about the pattern, but it does have mostly the right idea. I like the detail it has in some areas, but then there are other areas that just look airbrushed in and it makes me feel weird.
Next thing I told the AI to generate was “shorthair calico cat with green eyes, a pink nose, with a bigger build in the style of Mona Lisa,” which it responded with this:
And hey, that’s starting to look like my cat! But it’s not in Mona Lisa style, but that may be my fault because I still had it set as the “photograph” setting. So here’s me retrying without any settings:
Now I’m confused, because it’s like half Mona Lisa-esc but also just not at all. So I decided to play around with some of the model settings
This is the same prompt but in the Rococo model, and it’s actually really cute, I won’t lie. I like the random bow that appeared and the colors.
I did one more just to see what it would spit out:
This is in the Baroque style. I guess, of all of them, it did get the Mona Lisa style down the most, but not entirely. I’m sure I might just be putting in the wrong queue words or not being specific enough, but I think it’s cool.
My overall thoughts on this?
It’s cool, and fun to play around with, but not cool or fun enough to ignore the threat that it poses to artists.
I think it’s really, really threatening. Even if it doesn’t take from artists like people said, how can you prove that it doesn’t? It learns from inputs and it learns from taking images and prompts and anything you can find on the Internet to build it. It learns, and it implements what it learned into its model for future purposes. It feels relentless, like publishing any art on the Internet puts you at risk of having it taken and tweaked just so someone who knows how to type in a good prompt can get the same output. There isn’t creativity in prompt-making, but there is creativity in creating art.
It’s not art, it’s data, but the data is built off of previous ideas of artists, taking it and transforming it into some code to create more data.
I enjoyed what Digital Detox had to say about art and AI. Largely, I agree with a lot of their comments and their outlook on AI. I do think we have an obligation to protect human artists first and foremost instead of allowing everyone to turn to AI for their art needs. On the other hand, I do think AI is something we need to learn to live with and manage because it’s not going away anytime soon. I liked the comparison of our current situation to the invention of the camera and how some artists in the 19th century were also very skeptical of its status as art.
For my entry, I wanted to go with a very simple prompt of asking the AI to create the four changing seasons. I used Bing and Stable Diffusion. After this very simple prompt, I got these outputs:
Stable Diffusion
From this round, I think that Bing is the winner because Stable Diffusion is just showing the same season four times. For my improved prompt, I asked both of the AI to add people and I asked Stable Diffusion to make sure to show white for winter, green for summer, orange for fall, and flowers in the spring.
Stable Diffusion
I believe Bing has won again because it was able to show me the different seasons with people. Stable Diffusion still wasn’t able to show the differences in the seasons too well and it didn’t add people.
From the assignment bank, I decided to do the evil business card and warped text assignments. For the evil business card, I did one for my character because he is somewhat self obsessed and steals money from donors so I think he falls under the bad guy trope.
For the warped text assignment, I was supposed to warp the text of a poem around an object. Unfortunately, my editing software doesn’t give me the option to warp text, so I did a text overlay as it’s similar. I wanted to pick a poem with a flower theme to match the picture. The poem is called “The Flower Poem” by Alfred Lord Tennyson.
Share the image you felt was the worst. What did you ask for? How would you describe what you got?
For this image, I believe it is my worst because I had asked the AI bot for a cat having a tea party but the cat looks a little bit weird and I also have no clue what animal is next to the lady.
Share the image you generated that was most impressive. What made this one work?
This cat looks the most like my cat which was really impressive to me! I really love the accuracy that it has! I know I will show my cat in future assignments so you guys will be able to compare if you keep up with me!
Reflection: Where do you perceive AI image generation as a threat to artists?
I feel like the accuracy in this image can be a threat to some artist because of the details in it. the little hair strokes and eye crusties is what I think is accurate in art.
My other AI generated photos!
I had a theme with cats and I first looked up the breed of my cat and it expanded it to these images! The first image was also a fun prompt that I said cats having a tea party and cats riding a bike.
Assignment Bank!
these were my assignment banks! my first one was Your story as a game and Valentine card! I Used AI as the base and I added my own touch to make it match with what the design bank said! I hope everyone enjoys!
