Artificial intelligence is here, and it is not going away. It is reshaping the entire internet. In this changing environment, adaptation matters. I have been thinking about how myNoise can remain strong without losing its values. And about how AI is used, and should not be used, on this project.

I am genuinely fascinated by the way AI can assist my daily work. English is not my native language, and AI helps me write more clearly. I am coding the entire website, even though I have never been trained as a programmer. Here again, AI reviews my code, suggests improvements, and helps me implement ideas that I would have found too difficult to tackle on my own.

Refusing these tools for purely dogmatic reasons would ultimately harm the project, especially in a world where others will not hesitate to use them, and may even attempt to clone myNoise with little more than a prompt.

At the same time, I believe that the growing presence of AI will also create a counter-movement: a renewed desire for human-made work and for genuine connections between creators and their audience.

In that regard, myNoise has a strong foundation. I have never used generative AI to create sound for this project. The audio content on myNoise is entirely human-made, and free from generative AI. I never have used a prompt to create audio on this website.

I choose to continue on this path, and I commit to keeping the core of myNoise, its sound, human-created at the source.

This does not mean rejecting AI altogether. Refusing these tools at this stage would already negatively impact the project. AI helps me automate tasks that were slow to complete, and allows me to devote more of my time to creative tasks.

The sounds you listen to on myNoise are recorded or created by humans: by me, and occasionally by external contributors. This means field recordings made in real places, and musical soundscapes played or programmed by musicians. This is what gives the project its authenticity.

I do not use generative AI to fabricate the soundscapes themselves. Sound is the core of myNoise. If that were outsourced to AI, the project would lose its meaning.

Sometimes, randomness invites itself into my creative process. Certain instruments or systems may behave in unpredictable or procedural ways. I will keep using synthesizers for example, for their ability to generate random patterns or create ever evolving pads. This does not make them AI. They do not learn from data and certainly do not take over creative decisions.

I may process audio recordings using tools that rely on recent technological advances. Many modern audio tools already contain machine learning, sometimes labeled as AI. I consider this acceptable when the tool assists my work without creating content on its own. Noise reduction, audio restoration, or smart compression that reduces louder passages in recordings without introducing distortion all fall into this category.

The line becomes more subtle when machine learning is used to invent new timbres, such as the sound of an imaginary instrument. Technology has always been part of digital musical realm, and this is something I am fine with. I have always been passionate about creating sounds that have never been heard before. I can see a future where AI would be used to create a timbre, or even an instrument, which is then played by a musician through an electronic keyboard. This possible use of AI to generate music in combination of a human player, is fine with me.

But I will never ask an AI to “generate the sound of a forest,” “invent rain,” or “compose an ambience,” and then use that output to deceive my audience.

As part of my reflection, I have also been looking further ahead. What if I could train an AI on my own sounds to explore new kinds of transformations or remixes?

If that day ever comes, such experiments would be clearly and unmistakably labeled as AI-generated and kept separate from the core of myNoise, possibly under a distinct section of the website such as “myNoise AI.”

This case is purely hypothetical at this time of writing. It is stated here only to make one point clear: no AI-generated sound will ever appear on myNoise without users being fully and explicitly aware of it.

I may freely use AI for non-core activities, such as language correction, rewriting or structuring of text, coding, technical problem-solving, or other support tasks.

These uses do not require explicit disclosure. They do not alter the authorship of the sound, nor do they misrepresent the nature of the project. This manifesto itself serves as sufficient disclosure for these categories.

I understand that non-audio assets, such as visuals or illustrations, can be a sensitive topic. I am thinking in particular about the background images used for each sound generator, which take a considerable amount of time to select so they match the sound appropriately. I spend hours finding the right image, and sometimes such an image doesn't exist, neither in my own collection nor in the stock image services I use.

Human-made visuals are always preferred, especially for soundscapes based on field recordings. It would not make sense to pair genuine nature sounds with imagery that has been artificially generated. There are categories, however, such as synthesizer soundscapes, where computer-generated images may best illustrate what the sound is about. In such cases, AI image-generation tools may become handy as a fallback, only when no suitable human-made alternative has been found.

There is, however, an ethical boundary I am imposing on myself. AI must not be used to reproduce the identifiable style of other creators. It may only be used to support generic concepts, such as abstract themes, widely shared representations, or public-domain knowledge, without evoking the work of specific artists.

To clarify, here's a concrete example from myNoise. The soundscape called Klang Machines is a nod to Kraftwerk. All the audio tracks are my own, recorded with real analog synthesizers. When I prepared a background image for that generator, I wanted something that evoked the band without copying their work. I used AI with a prompt like 'robots made of kraft paper representing the members of Kraftwerk.' For me, this use is acceptable. The prompt mentions Kraftwerk as public cultural figures, not as a style to imitate. The image doesn't borrow from their visual identity or replicate any artist's recognizable style: texturing the robots in kraft paper does not seem something that is the trademark of any known artist.

This is an edge case. 99% of the time, I can find human-made images that fit a generator perfectly. AI only comes in when the idea is unusual, when no suitable image exists, and when I can frame the prompt in a way that does not borrow from any artist’s style.

In a future shaped by AI, trust will matter more than ever. People will look for places where they know what they are listening to, who made it, and how it was created.

myNoise is built on that trust. By keeping sound creation human at its source, and by being transparent about the tools I use, I want to preserve a respectful relationship with my audience.

Some may feel that the position outlined on this page does not go far enough, and that any use of AI should be rejected outright. I understand that reaction. It often comes from a deep concern for human creativity and authenticity. But for a small, independent project like mine, refusing all such tools is not an option anymore. It would only weaken a fragile creator, while larger players move forward without hesitation.

AI helps me regain time from repetitive or technical tasks. It does not take creativity away; it allows me to allocate more time on what matters most: designing original sounds and features for this project.

I hope this approach will be understood. And if it is not, I hope at least that my intention is clear: to be transparent with you, my dear users, in a world that is changing more and more rapidly.

Published on Jan 6th, 2026