We are an international team of professionals in love with music and our goal is to provide top-quality services for musicians and music bands all over the world

AI-generated Art for Music Bands: Pros and Cons

AI is on hype now. It's transforming every aspect of our economic and social lives. AI models find applications in search engines, chatbots, speech recognition systems, and even autonomous vehicles. Moreover, they also generate images, videos, and even music. Chances are you've already encountered numerous AI-generated images used as album cover art, YouTube visualizers, and concert posters. In this article, I will review the most common AI models for image and video generation, explain their capabilities, and delve into the pros and cons of using AI art within the realm of music bands and music-related projects.


What is AI art?

AI art, also known as generative art, refers to artwork created using artificial intelligence. This art form is generated by inputting data, such as text, images, or sounds, into an algorithm trained to produce new and original artwork. Already, thousands of AI systems are in use across various industries and applications. In the realm of visual art, the most common AI models are those capable of generating images and video clips. I will now specify the three that, in my opinion, are the most commonly used or referred to.

Midjourney is an AI model that generates images based on a text and/or image prompt. Midjourney’s style is the most “artistic” among the other AI-generators. It uses Discord servers and allows the generation of images via cloud technologies. This means you don’t need to have a powerful PC to work with it. You can simply connect to the bot via Discord and start feeding prompts to it. Midjourney is not completely free. They allow you to try it through a free trial, but then you must pay a subscription fee that starts at $10 (basic plan). In most cases, the “standard plan” is what you are really looking for and it costs $30 monthly.

Midjourney example


Stable Diffusion is another text-to-image model released in 2022. Its main difference from the other AI-art generators is that it is an open source project meaning everyone can download the code and use it on their own PC or a server. This has some benefits for the end users. It's free, it can be run on your own PC (if it meets the requirements), and you don't need to worry about censorship that the other AI models have. If you don't have a powerful PC, the image generation may be quite slow. There are servers on Discord and Telegram that allow you to generate images via cloud using a chatbot, but these are more limited. In my opinion, images generated through Stable Diffusion are a bit less realistic and “artistic” compared to the ones generated with Midjourney. However, SD is much more accessible.

Stable Diffusion example


Kaiber.ai is an AI model that generates animated videos. It costs from $5 to $25 monthly depending on your needs. The most popular plan costs $10 per month. Kaiber.ai generates animated visuals based on your text prompt and it also has a feature called “audioreactivity”, which makes the animation respond to a music beat. This sounds pretty cool, but to be honest, videos made by this AI are usually quite chaotic and hard to follow. If you want to see an example of an animated video made with Kaiber.ai, check out THIS VIDEO released by Linkin Park. During 2022-2023 they released a bunch of demos and unreleased songs with visuals made by AI.

If you want to check out other AI models for image and video generation, you can check out DALL-E, Dream, Craiyon


Pros and cons of AI-generated art

The pros of AI-generated art are obvious. It’s cheap and it’s easy to make. You choose an AI model you like, pay a small subscription fee, spend a day learning the process, and then you are ready to generate as many images or videos as you like (within your subscription plan). AI tools significantly reduce the costs for products that require visuals without high expectations. Like a gig poster for a local metal concert or a Spotify playlist cover image.

The cons of the use of AI art are less obvious for many, but they are quite serious.
1. First, there is an ethical problem. To “teach” an AI to generate art, developers usually use the art of millions of graphic artists and photographers on the internet without their permission. Many artists from all around the world have protested such a method of AI learning.
2. Secondly, there is no easy way to make sure the image that AI-generated is original. Sometimes it makes distorted versions of existing art or combines bits and pieces from it to create something new. There is always a chance of copyright infringement claims, especially if you use AI-generated art for commercial purposes.
3. At this stage of AI development, any person with at least some background in graphic design and art can say if an image is made with AI. AI makes mistakes, blends objects in a weird way, doesn’t know anatomy, assumes proportions wrong, and often has problems with depth of field. Some things may not be obvious to an average user, but they are obvious to any artist/designer and sometimes make them cringe.
4. AI often repeats itself. Similar images can be generated by different people using the same prompt. Especially, if AI had limited "teaching materials" of the requested topic.
5. At this moment, the perception of AI art by the public is not very good. If you use AI-generated art for your band’s album cover, for example, people may think you don’t take your music seriously and that it’s not a quality product.
6. AI lacks creativity and an artistic approach. All it does is replicate the most obvious visualization of your text prompt based on what it “learned” from the internet. This kind of art is very “sterile” and doesn’t have any depth. Moreover, if you are not a trained artist and don’t have any experience in design, your perception of AI-generated art can be very subjective. Something that you personally think looks good may be perceived by other people as bland and uninteresting


Generation of final product vs generation of assets

When people talk about the use of AI-generated images, they usually mean the final product. However, some artists and designers generate bits and pieces that they either create original art from or use in their own artwork. This is, in my opinion, the most appropriate way of using AI-generated art. For example, we use Stable Diffusion to make silhouettes of creatures and people, which we can use for our digital painting and matte painting pieces. We use Midjourney to generate landscapes and architecture if we want to incorporate it in the far background where it’s heavily blurred. And we also generate textures sometimes. Such solutions help us make original artwork more time-effectively.


Are we against AI art in general?

No, we are not. We as a studio feel that despite all the cons, AI art still has its place in the industry. There is no point in denying the progress of AI models. They get better and they probably won’t go anywhere, so we must adapt. Hopefully, they will be regulated better by both the developers and the governments, so artists all around the world won’t feel as threatened by AI as today. Of course, artists will lose some part of their income because of the AI. Especially the ones who don’t have a recognizable art style and good skills. The artists from the lower-price segment will be affected even more. Still, we think the demand for visual artists and graphic designers won’t be significantly reduced. People with real skills and talent will always be appreciated. There are situations where the use of AI art is suitable, but there are a lot of situations where it is harmful and inappropriate.


Inappropriate use of AI-generated art

First and foremost, I don’t recommend using AI-generated images as cover art for your music, unless it’s a very clever use of AI art or if it’s edited by a professional artist/designer. If it’s obvious that your cover art is an AI-generated image (and maybe not even a good one), the overall quality perception of your music may suffer. As I mentioned above, AI art is usually associated with low-budget and low-quality projects. There are accomplished artists that have used AI-generated art for their music releases (Currents, August Burns Red, Cradle of Filth), but it’s a phase. Moreover, artists who already have a name can use AI art more safely because their music is already in high demand. Young artists usually don’t have this luxury. By not taking your album art seriously you miss an opportunity to make your product more eye-catching and desirable.

Secondly, I don’t recommend using AI art for any commercial projects. By commercial, I mean anything that is intended to generate money (a music album, a big curated Spotify playlist, monetized YouTube content, any kind of merch). A high risk of copyright infringement claims isn’t worth the money saved by the use of AI-generated art.

And lastly, I personally don’t recommend using AI to make concept art and reference images when you order something from real artists. In my experience, clients rarely understand the difference between different techniques and their limitations. By making images with AI, they set expectations in their heads that may be inadequate or misleading. A piece of AI art generated in 1 minute may give the impression that the artist should easily make “something of the same level but better” with little effort. AI art may be weird, and it has a lot of flaws, but the level of detail that it usually has would be very difficult to achieve drawing by hand. On the other hand, by providing AI-generated reference pictures you also limit the artist’s vision. It’s much more helpful if you use the artist’s own work as a reference for the style or composition that you want.

Appropriate use of AI-generated art

In my opinion, it’s okay to use AI art as a final product for non-commercial projects where you simply don’t have a budget at all. It’s also okay to use AI for images, that are not essential for your branding. Sometimes spending money on something that most people don’t even notice seems wasteful. In my opinion, these are good examples of appropriate uses of AI-generated art as a final product:
- FB and IG ad creatives for low-budget or non-commercial projects.
- Spotify playlist covers, FB cover images.
- Gig posters and flyers for underground concerts with low budget.
- Cover art for hobby music projects that are completely non-commercial.
- Art projects where the use of AI-generated art is conceptual.

Hopefully, I have covered most of the common questions about AI art and its place in the music-related design industry. If you want to leave a question or a comment, you can do so on our FB or IG page.

Vladyslav Tsarenko | CEO at All4band Design