Will Artificial Intelligence replace artists in 2023 (Stable Diffusion, Midjourney, GPT3, DALLE etc)

preview_player
Показать описание
In this YouTube video, "Will Artificial Intelligence replace artists?", Allan McKay discusses the potential impact of artificial intelligence on the art industry. As AI techniques like GPT, GAN, and diffusion models continue to evolve, many artists and professionals are wondering if machines will eventually replace them in their careers.

You might be familiar with Stable Diffusion, Dream Studio, Dreambooth and DALLE, Midjourney and of course GPT3 and chat GPT all coming out in the past few months and taking the world by storm with their advanced artificial intelligence!

Allan explores the ethical considerations of AI in the art world, as well as the current state of AI in industries like visual effects. He also offers insight and advice for artists looking to navigate this rapidly changing landscape and maintain their careers in the face of increasing automation.

Whether you're a 3D artist using software like 3ds Max, Maya, or Houdini, or a VFX professional wondering about the future of your industry, this video is a must-watch for anyone interested in the intersection of art and technology. Allan covers the potential of AI to both enhance and disrupt the art world, and offers tips for artists looking to stay competitive in a world where machines are constantly improving their ability to create art and design.

Chapters:
00:00 - AI Revolutionizing the Visual Effects Industry
01:34 - Embracing AI in the Visual Effects Industry
02:54 - Communicating with Computers: From Assembly to GPT
04:16 - Communicating Directly with Artificial Intelligence
05:42 - Exploring the Possibilities of GANs and GPT
07:12 - AI's Fearless Pursuit of Improvement
08:21 - AI and Generative Adversarial Networks
09:40 - Training Machines with Trillions of Parameters
11:06 - How AI Affects Artists
12:48 - The Impact of Bell's Palsy on Performance Capture
14:32 - Disruptive Innovations in the Art Industry
16:06 - The Fear of New Technology
17:20 - Mass Innovation in Computer Art
18:41 - AI-Driven Innovations in Concept Art
20:05 - The Possibilities of GPTChat and Deep Fake Software
21:30 - AI Innovation Far Ahead of Expectations
22:42 - AI Replacing Artists: Is it Inevitable?
24:29 - The Fear of New Technology
26:11 - Adapt or Die: The Need to Adapt to Innovation
27:50 - Adapt or Die: Staying Ahead of the Curve
29:25 - The Impact of Cheap Content Creation
30:58 - The Impact of AI on Copywriting
32:27 - AI Generated Art and Watermarking
34:10 - Adapt or Die: The Impact of Automation on the Creative Industry
35:50 - The Power of the Art Director to AI
37:29 - Executing the Director's Vision with AI
39:10 - Leveraging AI Tools to Create Art
40:48 - Leveraging AI in the Art Industry
42:28 - Embracing Innovation with AI
44:02 - Exploring the Business Side of Art

Playlist:

Please sub?

Listen to ALLAN McKAY’S PODCAST:

Upload THE ULTIMATE DEMO REEL GUIDE by Allan McKay:

Upload PRODUCTIVE ARTIST GUIDEBOOK by Allan McKay:

SAY HELLO ON SOCIAL MEDIA:

YOUTUBE URL:

VIDEO HASHTAGS:
#AllanMcKay, #VFX, #StableDiffusion
Рекомендации по теме
Комментарии
Автор

I am a concept artist, and way I see what AI will do to my industry is, it will drastically lower the time of producing 1 image. I think that's where people are getting angry about. But my job was never really creating image itself, it was more about create images to solve the problems. So if using more images with lower cost will going to lead me to solve more problems, even on that case, it doesn't affect me because there're endless numbers of problems on any projects to solve. But... usually, more image doesn't really mean solving more problems, because "choosing one from multiple variations" was never really good method of solving problems in my experience. And even if AI will going to create images so fast, what is the point on that when I am not solving problems alone? It involves so many people and departments on each problems, and me creating images like madman will not really gonna get me anywhere. So... yah. I think I am gonna be fine.

seokinchung
Автор

I'd love to hear your thoughts on the impact of AI in the VFX industry! Have you seen it change the way you create art? Share your experiences and predictions for the future in the comments below. Let's start a conversation on the exciting developments of "The Future of Art: How AI is Disrupting the VFX Industry." Don't forget to like and share with your friends and colleagues!

AllanMcKay
Автор

Damn, you’d hit every point I’d been thinking about in this video and your podcast. It’s a reason I hadn’t done a video about this, and just threw up a poll.

ML is amazing tech. Scraping work without permission is a huge issue. LAION-5b non-profit turned profit at the potential expense of artists is an issue that the courts will take up. Artists have a right to fight, though I do think the hype will die down, we will see tons of crap on the net, and once artists adopt this tech, they will steamroll over the average user. ML in the hands of trained artists will in many ways allow artists to own the future.
As a generalist mostly using Houdini at work (Blender is my YT thing, ) I have a feeling SideFX will do something with this tech that will blow the lid off everything. Just my crappy opinion, but I’m holding to it lol. Happy New Year!

KevBinge
Автор

Hi Allan. Greetings from Finland! Thank you for this super important video. I just hope that as long as a real/human artists are in charge of polishing the final product, then everything is fine. The real artists with the decades of experience will be always in high demand.

georgelevantovskij
Автор

A simple story of a hypothetical event:
you spend six months learning some skill.
you take a job interview that asks you to demonstate that skill
you are recorded in the job interview, doing the skill
the recording turns out to be sufficient information for a computer to replicate the skill
you are not asked to come for a second interview

davib
Автор

Thanks Allan! I personally needed to hear this. Been following you since 2006, even took your Fume Fx course and Pyro for Hollywood Fx on CGtalk back in the day. We even chatted a few times! So I've always taken your advice as artist and will do so today as well. Cheers!

thetoweroftomes
Автор

James Cameron cannot yell at the Ai, 😂 that comment made me literally laugh so much in a positive way, but also made me truly understand that Ai won't replace VFX and the human factor, but just change the way things are made and so for me to continue pursuing my passion even in my 44 years old😅. It's like when computer graphics came out, artists who jumped on board the boat and embraced it evolved, whereas others were left on the shore. Thank you so much for the, as always, inspiring video! Keep the great work!!

nikolaostsimpetonidis
Автор

I completely agree with everything you said! I have been looking at these new technologies as tools to help make my job easier / faster, regardless of the task. The generative AIs can produce some amazing results on their own but the output from the AI is rarely a final solution, and will always require human input. Or when it is a "final solution" it's very much like the example of trying multiple filter generations until you get something that you like (I'm guilty of that for procedural textures, which these image generators can now help a lot with).
Also, one thing you probably should have mentioned, is how easy it now is to train "custom" models. So for example, if you want to generate really good concrete textures... you can fine-tune a SD model with a bunch of photos of concrete. And that model is one that only you have access to (unless you share it), so that can help give artists / studios an edge. The same thing can apply for concept art.
Another really good use would be visualization of ideas before execution. There have been many times where I tried to flush out an idea in my head, spending hours working on it, and the result was lack-luster. Being able to short-circuit that process without spending all of those hours can help guide early artistic decision making (that may not be as much of a problem for someone with a lot of experience since they have developed strong intuition, but it can really help junior artists).
And of course, it helps people who are not professional artists produce things that they otherwise couldn't. The quality and direction might not be as good if a human did it, but in many cases it's "good enough" for personal / low-budget use.
Taking that last idea to the extreme, there was a game recently released which used a GPT style model to generate a text based story and interaction, and stable diffusion to generate the images (from the GPT output). While the quality isn't great, each interaction is custom and unique to the player (something that would not be feasible if artists were to produce the assets - images, stories, interactions, etc.).

As for the comment about watermarking, that's been in the official SD repo since day one. Though I think some of the web interface tools have removed it... perhaps the contact at Stability meant embedding the watermark into the model weights so that it can't be removed? (not sure if that's possible with the VAE)
And digitally tracing the source of a generation is not feasible (I'm not sure why people keep suggesting it). If it were possible, then it would be a good solution, but you can't pull out what image adjusted which of the weights. In fact, if it were possible, there are far more interesting applications, such as adjusting individual source contributions per generation. The best you could possibly do, is to charge when an artist's name is used in a prompt, but that's not possible to enforce with open models like stable diffusion.

hjups
Автор

I'm surprised AI isn't moving way faster in areas like unwrapping UV's and retopology. Seems like an area where artists and companies could happily benefit together...

charlieharris
Автор

Hi Allan Thank you for sharing this video. I have seen a huge opportunity in the AI space as well. I even left my job in order to pursue motion graphics/VFX in the midst of all this! But I am glad we can at least leverage this new tool for our toolboxes!

jonathanabramovich
Автор

Very small freelancing artist here. I'm both fascinated and terrified with how fast AI is sweeping over. I'm still to notice any changes to my livelihood, but this video has been very enlightening on what optics I should otherwise exercise, what I should focus on and how the general profile of my clients will change or endure. I'm still fumbling to adapt, but having someone experienced and involved like you go over this really helps dissipate the fear and uncertainty that has been hitting me in waves these last months.

jaegermonster
Автор

The real problem with the AI vs artist debate, at least as far as I discussed it with various professionals across the industry - people still think that an image is the actual product, not the thing this exact image is for. They are afraid of losing their jobs because they see their work to be producing images - paintings, concept arts, sets, you name it. And that's where AI is (or will be) better, or at a very least hella faster.

The way I see it, from an art dir perspective (but also with my hands being in concept, ui and all sort of graphic stuff personally) - no one ever hired concept artist, or an illustrator, to make a bunch if images. The goal was always to produce and add a value, or procure a certain vision, to the broader product so it could have been understood and sold to the audience. Let's take the concept art - the job of CA is to come up with the designs taken further into pipeline to make them into assets, that will later end up as a game or animation or whatnot. Let's take an illustrator - the job is to visually describe a certain, more or less abstract and hidden withing a book/game, idea so the audience could see it in a certain way, enhancing storytelling of the said product. It's never about creating images for the sake of images.

So the most important thing that will have to happen in the industry is the pros to understand that their images, styles and technical mastery of their craft is really cool to have around and all that - but it was never the main value in the industry to begin with. It will be painful and for some a bitter pill to swallow.

adamdawidowicz
Автор

Well thank goodness Allan put out a video taking some sense into the fear mongers all over the internet in the last few months. Thanks for sharing your thoughts.

joelpaulin
Автор

Of course this video was going to be one of those that talks common sense for AI and future of art/artists, since you're a very experienced one. Thank you.

arbitool
Автор

I think the embedded metadata idea is a good one sir, as well as many other points you made

olajiderasak
Автор

Thanks Allan, I have been in the industry for a while and been testing AI and for me it's a great tool to use to simply my task . If you work on big production like Avatar or any big blockbuster they are so picky on every detail that AI will be only great to get the shot up and running faster, but good luck finishing to the level they want with AI . Hopefully it will help reduce the insane amount of OT we do on those big production.😉 A good artist will always have is place, but time will tell us.

YTFPV
Автор

I would use AI as a sweatshop slave labor to get most of my monotonous work done so that I could be as much of a director as possible in my field of work. This would mean basically for any individual in whatever line of work (in this case artistic industry) to spend more and more time developing an Eye for the aesthetic. You will know what something should be a standard and improve your capability of presenting that versus individually developing those micro skills necessary to deliver a highly polished piece of work. Not to mention the possibilities and growth of the individual.

It's similar to farming agriculture and how it freed majority to focus on other things in life now that they freed their time. In turn we were able to achieve more and be where we are today. I think AI will make us better artists/creatives and worse manual craftsmen if not completely redundant to begin with. I doubt classical art will ever go away, it will be something like how having and riding horses today is delegated to the cultured and the rich or an expensive hobby/passion with it's own community. So long as humans exists, they will always value humanity over generated drone work. I think AI would actually possibly rebirth a window in time again where art will be valued again as it will lose it's abundance that is currently present due to demand and lack of technology to meet it.

blah
Автор

Hi Allan, I see you are recovering. Godspeed!. I just listened to you saying about Bell's Palsy.
Please don't find this question odd, , but I like to listen to everybody when I am concerned so I can study and understand better about it. question is that did you take the covid jabs ?

saadomar
Автор

There are many aspects of VFX that are daunting to do or time consuming such as rotoskoping but there are people who are enjoying the artistry of it and AI is taking those jobs away. I'm totally perplexed why there are people advertising or even working on the code to help AI advence and take even more jobs in the VFX hierarchy. Let alone the massive compyright issues on creating AI 'art' which must be made illegal!
Well you're not going to keep learning if a technology comes out and takes the whole industry away from what you can do, in fact then you'll become stale and inactive as something else will be doing it for you.
I believe assisting the creation of AI code is immoral!!

No_Plastic
Автор

I'm literally up late because I couldn't stop thinking about this.

I've never coded in my career, and this week using ChatGPT I'm busy writing a pretty robust shader converter. It's wild!

Thanks for the video!!

pieterlouw