Skip to main content

The Use of AI in Our Work

Do we use AI (Artificial Intelligence) in all of our work?

The simple answer is… no… but also yes. We believe that technology should only be part of the process, not the entire process. This goes the same with the use of Artificial Intelligence.

AI can be a very controversial topic which offends or upsets a minority of people but we do not plan to use AI to create 100% AI generated content.

Instead of using AI to generate complete works such as for example an Anime style drawing of a game character, we could use the AI generated art to gain an understanding, the concepts and inspiration for creating our own original art based on it.

Many people make complaints about Al’s use of so called “stolen art work” but you know… those same people already stole other peoples works in the same way when they learnt to draw; by looking at Google images, TV shows/movies, magazines, books, teachers, games, fan art and so on.

Also, whether or not you choose to believe it but redrawing someone else’s work is an infringement or their IP (Intellectual Property). How many people have drawn art based on something they’ve seen created by someone else? Pretty much just about everyone has, please don’t deny it!!

Imagine, you wanted to draw fan art (“Stolen IP”) Yae Miko from Genshin Impact wearing a particular kind of dress in a certain place. However, you’re not too sure how to even start it or what it may look like (hard to visualise it in your head). You could gain some inspiration and ideas from Al generated art and then pick the ones you like and use those ideas in your own original art.

“Fan art isn’t protected by copyright” – Actually, it is; copyright exists the moment something is created, it is not required to register that work. Also, you are not allowed to enter even your own works into the Public Domain; this is why the Creative Commons came to exist but that doesn’t mean it has no copyright, it is still owned until it expires based on Public Domain laws.

※ Fan art, fan music and so on are illegal because they are derivative works of protected content but stopping it is impossible to enforce and for that reason many companies and individuals just accept it as long as it doesn’t harm their reputation or cause them other legal issues.

How about AI in our music recordings?

Again, while we do use AI, it is only part of the process. AI allows us to speed up the process of creating a recording by understanding and hearing the sounds and then manually editing the music to sound better than what the AI generated.

For example, pushing an entire orchestral score into a DAW, then using VST samples to play everything back sounds static, unrealistic, boring, artificial, flat and certainly nothing like a real orchestra. It’s also difficult to hear how each instrument and section works together.

When conducting a real orchestra, this process is easier because each instrument is played by a real human who is performing in their own way but together as one. The conductor can then alter things and change the music in ways that suits the performance best.

With that, we use AI to generate a recording to give us an understand how each part may work together as if it were played by real people, then use this recording in a DAW to make fine adjustments using VST instruments and samples.

This includes a certain amount of creativity and uniqueness that are not mere copies of the AI generated content and for that reason is not 100% AI generated but is our own original work, copyrighted by us.

Is anything performed live?

Yes, there certainly is. Most piano recordings and even piano parts of orchestral scores are played live. Why not? I mean, I didn’t spend over 20 years learning and studying to play the piano and getting all those qualification for it all to go to waste on some AI recording, especially when my own playing will sound a lot better.

Also, in regards to vocals, I will use AI, however, again, it is not 100% AI generated; that would sound pretty emotionless and robotic. AI even today does not understand how a song should be sung and with what emotion or character or style.

Creating vocals from AI requires a lot of work to manually fix. A whole lot more than just singing it yourself or using an AI voice changer. You must sequence the lyrics and the notes individually along with velocity, dynamics, tone, pitch shift, effects and so on.

Also, I really do hate my own voice and cannot sing too well. Is it really that wrong to sequence singing in a voice I like the sound of, in away that is expressive and performed just the way I want for my own work? especially if I am not loaded with money to throw away to hire a professional singer, who may not be available or disagree with parts or change parts to fit their own political agendas, even though I want to sing it, in my own way and faithfully?

You may not like it or even hate the idea but is there truly anything wrong with wanting to sing with the voice of the opposite sex? (i.e: you’re male and want to sing with a female voice / you’re female wanting to sound like a male).