 So before we dive back into Unreal I wanted to cover a short little series teaching you guys how to create your own data set using your own art that you own 100% of the rights to in order to train your own personal AI to give you exactly what you need how you need it and when you need it because I realized something important as I was trying out all the different AI. You know when we were learning Dolly it was really cool and fun but the list of banned words made it unusable for anything outside of cute generic pictures words that you often need in a video game like war or battle or fire or shoot would get your account striked and even certain styles like Quentin Tarantino would get your account flagged or shut down and even after moving to Adobe Firefly while it has been excellent at doing general things like abstract art and generative fill it's proven to be very unhelpful as soon as you need anything specific for example I have two friends that have wildly different art styles one of them is a horror artist and in horror you commonly need to explore themes like brutality blood gore torture mutilation you know the usual and Firefly obviously does not let you go into that stuff at all and my other friend is the polar opposite where they just love to generate cute anime waifus just a typical you know anime girls and every time you try and get that out of Firefly well let's just say it doesn't really give you what you would traditionally expect a typical anime waifu to be and I realized that this issue is never going to change as long as the tool you use is dependent filtered and maintained by someone else you're never really going to get exactly what you need as long as you're using someone else's AI what you can and cannot generate will be controlled by the person running that website and that's why I think it's so important that you have the ability to generate whatever it is you need or want not what people say you should want not what people allow you to want what you actually want whatever that is the only way you're really going to get it is if you train it yourself and that's why I ultimately believe that in the future everyone's just going to have their own AI trained on whatever they're interested in if you're a writer then your AI will be trained on all your favorite books if you're a programmer your AI will be trained in all the languages that you specialize in if you're a musician your AI will be trained on your favorite songs and likewise if you're a visual artist your AI will be trained on all your favorite historical artists your own works and your own style if you're into horror your AI will specialize in horror if you're into waifus your AI will specialize in waifus and this stuff is getting more advanced and more accessible every day and one of the reasons I decided to do the series in the first place is because I was recently shown you can train an AI reliably with about you know three to five hundred images which boggled my mind because I think most artists like myself easily have three to five hundred images or screenshots of our work lying around so when I realized this was an option I was like holy shit I could save so much time coming up with my own designs if I just trained the AI on my own work and honestly it's been a lot of fun because normally every day I wake up I have about a hundred new ideas for cool character designs but it takes me about three months to make one and sometimes you spend three months modeling a character only to realize man this really doesn't look as good as I imagined but now I can easily brainstorm and filter through all of my ideas to find the ones that I think the players are really going to enjoy the most and I figure I'm not the only one who would probably find this workflow extremely useful so I wanted to make sure to share it with you guys before we move on to our next project now something we will be doing that is a little different from most of the other AI training videos is we are going to be approaching this from a 3d character artist perspective which is going to help save us a lot of time because normally you have to spend a lot of time trying to generate images that look like the same character in different poses and lighting and it's very hard to keep things consistent across all generations so that process alone often takes countless of hours of frustration and often requires you to go in by hand and manually fix things like the face and clothing to make sure they're the same across all images but as a 3d artist we don't have that problem we have a rigged 3d model that we can control the lighting background and pose however we want and the character will always stay perfectly consistent in every instance by default so if you are a 3d artist interested in training your own AI on your own work then just join me in the next few videos and I'll walk you through step by step it's a lot of fun so I hope you're excited but anyways thanks for watching and as always hope you have a fantastic day and I'll see you around