15 INSANE Use Cases With ChatGPT 4o Image Generation

So ChatGPT just released the 40 image mode. And in that mode, there are a variety of different use cases that most people still haven't realized. In this video, I'm going to go over 15 different use cases that you can actually use today to start utilizing the new image generation. So without wasting any more time, let's jump right in. One of the first ones that you can actually use is combining styles. You can see here this user said, this is a mood board of how our image style usually looks like. Can you use it as a reference and draw an entirely new image of a computer man wearing a robe? So essentially what you can do here is you can combine multiple different art styles all into one, and you can essentially use this to create new styles of images. Oftentimes when you're trying to create a unique brand identity, sometimes what will happen is you struggle to do that even when looking at multiple images. But with ChatGPT's intellect being able to use all of these assets, you can, you can combine these styles. Now, this is something that is really useful, and I think it's probably going to be even more useful in the future because unique styles are going to be very hard to come by when anyone can copy any style. So I would definitely use this tool. Coming in at number two is, of course, Ad Creatives. This is essentially where you can easily create marketing assets for any major brand using simple prompts. For example, the user here says, generate an image, reproduce this Budweiser ad, but use the Boom Boom product image attached, change the copy to match Boom boom and have the subhead say free your nose. So you can see right here. What you're able to do with the Ad Creatives is you're able to take an existing ad template, then an existing product, and then you're able to combine those to create an entirely new piece of artwork. This is really useful for those who are not technically savvy and those who don't have the ability to think of new ad designs. This is something that allows you to quickly Photoshop your products into another ad creative, and this is allowing you to see how your product looks in certain environments. If you've ever wanted to test something quickly and rapidly, this is going to be the way. Now, another thing that you can do if you're trying to test advertisement creatives, and what I mean by that is essentially variations of a product placement is you can actually do that with actors. Now, we could have done this before using a variety of different applications, but. But now using simple images of our product we are able to generate a new character that is wearing the same product. One of the highlight features of ChatGPT's new 4o image is that it manages to consistently have the same object throughout multiple different images. Meaning that once you put one image in, the AI is actually going to be able to consistently be able to take pictures in a variety of different circumstances. So when the user says generate an image, use the Clairv product from the images attached to produce a photorealistic ad of a model walking through Paris carrying the bag and looking chick. The model should be dressed in a certain style. It's able to immediately replicate that along with the product. This is something that is going to save people a lot of time rather than having a specific product, having to do long and extensive photo shoots and having to spend tons of time getting, getting the right lighting, getting the right angle, and getting various different things. One of the largest problems that happens when you're trying to do this is there might be someone in the background, you might have to cross the road. It could be a bit dangerous, there could be lighting differences, the weather might not be effective. Maybe the area that you want to go to is dangerous. Maybe you just don't even have money for travel right now and you are trying to bootstrap your company. This is going to be a complete game changer in terms of ad creatives and especially when you're trying to use models and individuals that are going to be wearing your product. Now, of course, once again, you can combine different things in order to create different ad creatives. Like I said before, this is something where someone combined the Ridge wallet right here. Another ad creative, again, is where you just take a random image and then you ask it to make it into an ad. So this is something you can do time and time again. Oftentimes one of the most popular things that people are doing is, is they are taking the product and then they are placing that product in a different image. And this is something that is very high ROI for individuals in the marketing space. You can see right here, you can take any image and any style and then combine them to create something completely new. I definitely use this and I'll make a full tutorial on this soon because there are a few tips and tricks that you're probably going to want to know. Now, another thing that has become really, really popular is the Studio Ghibli style transfer. So Studio Ghibli is a Japanese art style that is really popular right now and this is something that is gaining worldwide traction. The reason that People are doing this is because ChatGPT manages to correctly have the characters and consistently transfer that onto any new style. You can see that with these images right here, it manages to maintain the consistency of the original image and just simply change the style. So the reason that people like this is that it preserves the original identity of what we've already taken. That means it doesn't just create something new entirely in AI, but it enhances what we've already taken in real life. There have been many, many images of this happening online. I'm pretty sure if you have a Twitter account, you will likely have seen this by now. But I wanted to show you guys this because of course, some people may not have seen this. Now, style transfer isn't just for the Studio Ghibli style. You can transfer your styles between multiple different things. And what I mean by that is you can get really creative with the inputs and the outputs. So this, these inputs and outputs are quite interchangeable, meaning that if you have a standard image that you'd like to turn into a sketch or a sketchbook, you can transfer that. Or if you have a sketch that you'd like to turn into an image, you can do that too. This is something that is really cool because you can transfer different styles. Once again, like I've said, you can transfer different styles and also different perspectives. Something that this allows you to do is to take a standard image and then transfer it into a low poly isometric game screenshot. And, and you can see right here, this looks remarkably effective. So style transfer is really important, especially if you're working on a certain style of game, a certain style of character. This is something that is going to be remarkably effective if you want to maintain a coherent style within your choice of work. Now, something else that this can do that I was really surprised by is infographics. Infographics are things that are essentially giving you a decent amount of information in a nice picture. So it's something that you can save, something that you can print, something that you could even sell as a PDF or a poster that individuals could have for years to come. So all you would need to do is simply prompt ChatGPT with that information and usually it manages to get all of it right. Now, oftentimes with dense pieces of text, there are going to be some mistakes, so you may need to clean up the final image. But oftentimes you, it does get around 85% of the way there. So you can see right here, it has the brain. The lung is probably incorrect. The stomach is probably Incorrect. The intestine is right, a few of these are incorrect. But overall it does seem to be pretty decent in terms of how it manages to make the image entirely. So I do think that this is going to be something that a lot of people do use. And I actually used this to build the COVID for a recent PDF that I built and it actually performed exceedingly well. So this is something that can create infographics images that are really effective. And like I said before, it's something that is really, really good at doing this. Now, of course, when we look at other infographics, you can ask it to create a screenshot of a Wikipedia page and you'll be able to see all the information there. You can do this for a variety of different things. Perhaps you are a teacher who wants some little cards to be able to hand out in terms of the amount of information that you want. This is where you can get really creative. Now, I am going to quickly plug my community, but I did just drop a very interesting video on how you can actually make money with this because there are so many different use cases that I literally cannot include in this video. And I'm actually going to be revealing my full plan on how I'm going to get to 10k a month using this. And since I've started this, I'm already at $500 a month passively. So that's going to be something that you guys probably want to look out for. As I was saying, infographics are something that you can easily do. You can describe them in a multitude of ways, and it's really useful for just creating quick posts on social media that you can use on Twitter, Instagram. It's just really, really easy to do. And this is something that I definitely recommend you getting into. Now of that I've been able to do is I've been able to have on brand social media. So what I did was I simply took a picture of my Twitter profile, then I just said into ChatGPT. Can you create a Twitter image that is in the same style of my Twitter banner? And now you can see it managed to create something called the top six jobs that AI cannot replace. It did the perfect text, the perfect background, and the perfect color scheme. The reason this is really important is that when you're on social media, you. You want to maintain a solid brand image across all of your social media accounts and of course, your brand assets. That means your tweets, your YouTube videos, your introductions, your outros. And that is something that is often quite hard because Sometimes you might lose the font, you might not understand the color scheme, and you might not understand how to transfer things between different platforms. ChatGPT now makes it really easy to do that across any kind of platform that you want. And I was able to do this with one prompt. And you can see right here, I can now use this anytime I want to post something on social media. So if I want to post on Instagram, I want to post on YouTube. Maybe I need a slide for something. I can simply use this as my reference and then I can get this design. This is going to be really, really useful for those who want to be able to do that quickly. Now, another thing that you can do is you can have user interface styles. If you see something on social media or maybe on Behance or any kind of website that you do like the look of, you can simply copy that with any kind of prompt. The best thing that you can do is you can use any kind of reference and then you can say what you'd like it to be about. You can see here, the original image is here. And the person said, make a card like this, but make it about Italy. So they managed to change all of the assets and make it about Italy for their reference. One of the really cool things I like about this is that it even managed to get the background where you can actually see the Italian flag. So this was something that is really effective if you're trying to generate more assets for something that you need to do. Another thing that is really, really effective that it's able to do is it's able to render text really effectively. Text was something that AI models struggled with for a very long time, but now we're seeing these AI models are no longer struggling with large pieces of text. If you do have any kind of prompt, make sure you include not too much text, but enough text so that it doesn't ramble on its own or hallucinate certain facts. But you can see here that someone managed to make this. And if you actually pause the video and read all of this, it actually looks very, very effective. There's only maybe one or two very small mistakes that you wouldn't initially notice on your first glance, which means that this is incredible when it comes to producing images that humans really like the look of. So I would say if you're trying to create images that have a large amount of text, this can be done. Here's the OpenAI team actually explaining how text rendering works.

People tend to say a picture is worth a thousand words, but being able to also render a few words or symbols can carry thousands of pictures with a relatively simple prompt like Visualize, an infographic explaining Newton's prism experiment in great detail with a wide aspect ratio and a dark blue background. So this is an example where one we're going to rely on being able to render text in useful ways, combine it with visual elements that actually ground what this text about this experiment even means, and hopefully help students who are more visual learn both through language descriptions of a phenomenon, but also a visual imagination of what the experiment actually looks like. It's no longer just about making imaginary scenes that look aesthetic and things like that, but it's really about communicating and imagining and doing so at the same time.

Another thing that they have is they have character consistency. Now this is probably one of the most underrated features because for the longest time, character consistency was something that traditional AI image generators struggle to do. The reason for this is basically the fact that every time you generate a new image, it's often quite hard to reference the exact same style. And it's pulling from millions of different images, trying to recreate something that you've shown it. But here I'm not sure what method ChatGPT is using because I literally found this image on the Internet and you can see it managed to create this different pose of the same character. So I said make this character do press ups and it managed to do that. Now there are tons and tons more examples of actually having character consistency and I think it is probably one of the most underrated things that you probably should utilize. This now means that you can easily, easily generate stories, you can generate, you know, video games concepts. There's just a variety of different things that you can do that allow you to create coherent things over multiple images that are a lot more useful now. So I definitely take advantage of this.

One thing that I'm really excited about Imagen is the ability to keep consistency in characters. I'm David Medina or DMed and I work on multimodal. What I want to show is one of my favorite prompts, which is can you create a low poly penguin mage? Make it very, very low poly. Surprisingly, it's sometimes hard to get very good low poly outputs. It's not like other image generation models where it tries to generate something based on just the text. Instead it uses the large language model understanding of what does the user want? What is the intent. I also like some board games, so miniature like games. So what I'll do now is generate a miniature from this. So ideally we'll see A penguin that looks like this with the same staff and a hat. So can you make me a realistic miniature as if a professor professional made this and painted it. This is what I think excites me the most about imagen. The other image generation models will try to create literally what you said. But what's special about this is one, it'll keep the context of this character and then two, it'll understand what I'm trying to ask it for and generate very similar model but in a miniature realistic style. It infers what I want. I don't have to tell it every little detail. One other realistic thing we could do is can you make a crystal version of this with light reflecting and very realistic? Again, I'm just giving a very, very simple things. Normally this is not enough for other models to generate something very detailed, but the model understands what I'm asking for. It'll think what type of style it should have. So this ability to really understand what the character is and make edits and understand what the user wants, for me it's just an amazing capability.

Another really cool thing that you can do with the new image generator is that you can also take image assets from different things. So oftentimes we see a very good thumbnail and we wonder how we can actually use those assets or where those assets have come from. What it allows you to now do is essentially rip assets from certain images. Of course I do want to say that legality of this is unknown. Do make sure you have the rights to do that before you start taking assets from different images. But it's really cool if you have an image and maybe you don't have the PNG, you can literally ask ChatGPT to take out the shark for you to take out the image, maybe take out the water and you can see you can get the final image there. That is something that is really, really effective. This is something that I think is going to be really, really useful, especially for those who are designers. Because now rather than trying to crop it out and remove things, you can literally just ask the AI to do it for you.

My name is Jianfeng Wang. I'm a researcher in OpenAI working on multimodal. The way how to generate the transparent image is pretty intuitive, straightforward.

Now let's, let's quickly take a look at how OpenAI are using this to generate transparent layers. Transparent layers are essentially images that have no background. So when you import them, you import them by themselves and you can literally have them as stickers, place them as elements, and it just makes Compositing the image a whole lot easier.

My name is Jianfeng Wang. I'm a researcher in OpenAI working on multimodal. The way how to generate the transparent image is pretty intuitive, straightforward. Now let's give it a try. Let's say the content is QTP cartoonstere squared size. This is the prompt. And now let's see what happened. The model will take the input and tries to generate the image. Let's give it some time. So now let's see what we generated. This is the transparent puppy. Another application is to make it a sticker. Let's give it try. Yeah, let's make stickers. We can easily overlay the transparent image onto any kind of background. Now we can copy the sticker onto our laptop, and then we can paste it here, make it smaller so it can be easily blended with the background.

Can you make me a sticker?

Sure.

I'm a smart researcher wearing glasses in a blue shirt.

Okay, let's give it a try because. Yeah, with glasses. Oh, we got it. Yeah, I think it works pretty well. Hopefully people love it.

Now, once again, you can use this for thumbnail creation. If you ever have a thumbnail or image that you just want to swap out. Of course, like I said before, you know, using this on ad creatives, you can similarly do this on thumbnails. So someone's taken a popular MrBeast thumbnail and they've swapped them out with Elon Musk, and you can see it looks remarkably effective. So I do think that this is probably going to change the YouTube scene, as you may start to see a lot of thumbnails that are really, really effective. Now, another use case that is rather strange is the super realism that you see from these images. This is an image that I've looked at quite a lot of times, and for the life of me, there's not really anything that I can point out that would reveal that this is AI. Everything about this image looks realistic. There isn't one single thing that gives away that this might be AI. I've analyzed it again and again, and there truly is nothing. I mean, the color looks right, the grain, everything just looks like, you know, an Eastern European country or a European country, probably even the uk. And I mean, there's nothing that really tells me that this is AI, which leads me to believe the level of super realism from this new application is completely outstanding. And it's true. It is really, really, you know, wild what you can really do with this. So I would say for the super realism out there, I do believe that we've entered that area where a lot of images you might see. Now, you will have to probably wonder whether or not it is AI generated, because if images like this are AI, I'm truly not sure what I could miss. Now, I also saw this image of Trump. Now, there have been many different AI systems that can create super realistic images, but ChatGPT is on another level because the images that it creates are super high definition. Now, I don't have the ability to zoom in here because I haven't loaded the image indirectly, but what you do have is you do have the ability to zoom into all the creases and all the folds. And I will say that the realistic detail of these images is remarkably different. If you actually generate an image on Sora, the level of quality that you will see is top, top tier, probably better than anything that currently exists, which is Quite remarkable considering OpenAI just decided to leapfrog others in terms of the capabilities. Now, once again, we can see the detail here. It looks remarkably impressive. And one of the things that I've seen is that it doesn't have that AI tinge. I can't really describe it in terms of words, but if you've seen AI images or AI people, AI men, AI women, you'll know that there is a certain look to AI images. It looks realistic, but there is probably like a 5% tell that you can just about tell that it's AI with these images right here, I'm not really able to tell that it's AI. So it's quite unfortunate that even the best of us now are not going to be able to distinguish between reality. Only if perhaps you know about the context of certain images will you be able to gauge whether or not it is realistic for the picture to be taking place. I think that's probably going to be a bigger tell rather than the actual composition of the image itself. For example, take a look at this image. If I told you that this image wasn't AI, you, you probably would think that somehow actors have managed to put together a bear to be able to do this. But no, this was actually on the official Sora website as a generation. So, once again having crazy Snapchat pictures. I do think probably there might be a flood of new Snapchat pictures where individuals are creating crazy Snapchat pictures based on their environments that are just really funny or quite hilarious. I was actually staring at this one for quite some time because the grain and the small not errors, but just the way how the image looks a little Bit unprofessional, adds to the realism factor. Now, something else that is also a really nice use case is material transfer. So if you've ever wanted to transfer material from one thing to another, this is something that you can do. So you can say, okay, new task. Look at the image of the girl, look at the image of the Marvel marble material sphere, and transfer the marble material to the character making a marble statue. We call this one Material Transfer. So you can see right there, it managed to do this remarkably effectively. And this is something that would save 3D modelers hours of time, because if you ever wanted to do this before, it was certainly a really big headache. So I'm not sure what many use cases this is, but of course, like I said before, if you're creative, I'm sure you'll find a use case for this. Another one is, of course, muppeting people. I'm not sure why, but this one is rather hilarious, gaining a lot of steam, and it is super effective in terms of the quality of the images. And another one is, of course, the ability to convert sketches into thumbnails. So if you ever have something that you want to, you know, sort of create and you want to be able to visualize that with a simple sketch, you can now do that by having your sketch as an input image. And this is something that I think enables a lot more creativity. Oftentimes when we are trying to do this, we would have to hand these sketches off to graphic designers and tell them, okay, I want X, I want Z, I want Y, and I want it done by next week. But even if you have a basic level understanding of what you want, you can simply put that sketch in with a few rough pictures and then chatgpt will come out with a very, very perfect image that is probably quite to your liking. So this is exactly what you will get. And I think it's really important that you take advantage of this. Now let me know what you think about this tool. If you think it is good, if you think it is bad, are you going to use these use cases? And if you're intrigued on how you want to actually make money with this, don't forget to check out the tutorial I just posted in my community where I actually, in detail, show five key ways that people are making money with this. And I reveal my plan on how I'm actually going to use this to generate upwards of 10k per month.