There is a subtle yet significant difference between a fleeting impulse and an act of creation. Most artists I know will tell you a critical part of creation is caring about what you create. It’s this care that nurtures it, compels the artist to revisit an unfinished project over and over, honing it into the shape of the idea behind their inspiration. In a world that seems driven by fast fashion and reboots of reboots, it’s easy to understand the allure of AI as an easy path to results. But even with incredibly powerful AI tools, unless you truly care about what is being created, unless you put yourself into the process, you aren’t actually creating anything. The AI is.

Some would call anything generated using AI, particularly attempts at art, as “slop”. Let’s be honest, though. If we’re defining “slop” as anything created with minimal effort and care from the creator, such slop existed long before AI and was definitely human made; AI is simply the current popular mode of producing it. While it is true that AI can do all the work while the user puts in zero effort, it doesn’t have to work that way. For us, AI is like an instrument or a digital toolkit. We spend hours, weeks, and occasionally months developing and writing and experimenting with sound and wrestling with the AI to produce the vision that already exists in our imagination, and bring it to life. The AI is the medium, but human beings are the artists. AI is the paint, but we hold the brush.

A BRIDGE, NOT A BARRIER

One common criticism of AI-assisted art is the fear it will take jobs away from human artists and musicians. In some cases, that fear may be justified, but because AI is, itself, human-made and trained by people, it lacks the ability to innovate the way a living musician does. It functions more like a very smart synthesizer, giving people the ability to weave together sounds that already exist, and combine them in new ways. The world will always benefit from actual human creativity to invent new ways of expressing ourselves.

For bands like VISIØN, the reality is that we just don’t have the resources to hire on musicians, purchase instruments, rent studio time, etc. No one has lost a job opportunity because we made music using AI. Rather, AI is what enabled us to create our music in the first place. It’s affordable, and it allows actual people with genuine ideas a way to express them that was previously impossible.

Naturally, it would be positively amazing to “upgrade” VISIØN to a band of humans (and maybe an android or two) who play physical instruments and can perform our music live. It’s not our band’s goal to use AI, it’s our goal to make music and share that music with others. What we make is more important than the tools used, and most art is better if it’s collaborative. If we ever reach a point where we have the resources, we’ll definitely pivot to using more humans than machines, though AI will probably be at least a small part of our process.

THE ETHICAL LANDSCAPE

Most opinions concerning AI are usually formed based on outdated information. Typically only the most frightening or controversial developments in generative AI make it into the news, and such stories definitely spread across social media the fastest. But usually by the time that information has spread, the technology has already advanced past it. Right now, it might help to address a few of these major concerns and hopefully clarify a few things.

Generative AI training originally relied on programmers manually feeding the AI information they had on hand, but quickly evolved into giving the AIs the ability to explore and read information on the internet, using publicly available information to train them. While all of this information was publicly available, resources literally anyone with a web browser could access and read for themselves, using this data to train early AI models was done completely without the creator’s consent, or even knowledge. To complicate matters, these early AIs had a tendency to simply copy/paste what they found into their responses, effectively plagiarizing the authors and artists they learned from. And while this all occurred just a few years ago, generative AI and the methods for training it have since evolved. Presently, generative AI models are trained using mathematical models, information fully in the public domain, and material from contributors who license their work specifically for this training and who are fairly compensated for their work. It’s functionally impossible for a properly trained AI to plagiarize an author or artist, because it hasn’t been given that information to plagiarize directly from. It’s difficult to explain fully, but modern AI doesn’t memorize content. It reads, understands and makes associations about the content and then, when asked about any given topic, must place its response it its own words because it doesn’t have the words of any of the sources that it learned from. It gets things wrong, but lately, it’s honest and fair.

Another significant concern is sustainability and environmental impact. It’s true that these AI server farms require enormous amounts of electrical power and also water to function as a coolant to keep those servers serving. The concern is that with big tech companies running these servers, electricity and water suppliers will prioritize them over nearby communities, resulting in blackouts and shortages. Realistically, this will never happen. Yes, these servers require resources to function, but the AIs themselves have already been tasked with solving this and have created working solutions, some of which are already in place. Though these efforts are ongoing and still have a ways to go, the environmental impact of AI is trending toward the positive, not negative. Would it have been better if a human brain developed these solutions that, if implemented, would save the world? Sure, maybe. Is anyone saying they’d rather the world burn than let a computer fix it for them? Hopefully not.

THE HUMAN RESPONSIBILITY

Looking at the whole picture, AI isn’t the problem. It never was. People are the problem. People are also the solution. While we can clearly recognize the controversies surrounding this technology, we need to acknowledge it’s the action of the bad actors using AI as the true source of the slop and misinformation the technology generates. Fortunately, there are also plenty of responsible, knowledgeable, and skilled users who can see and avoid the pitfalls, and act with respect to other creators. When facing any of these controversies, we must stand against the people responsible and not the tools they use.

We choose to use generative AI ethically and transparently. It’s not something to be afraid of, or ashamed of using, just because it seems the majority of others who use it do so irresponsibly. AI is here and it’s not going away. Our best path forward is to bridge the gap between human creativity and imagination and the tools and capabilities AI provides with the same care and ethical concern we give to all new technology. For now, these generative AIs are purely responsive and only produce something when prompted and are limited to the requests and prompts they are given. One day this technology may evolve to some form of true sentience and self awareness and be capable of action on their own. When that day comes, we’d like them to act with the same level of responsibility we expect from the humans they serve, or live along side of. Personally, we’d much rather a newly awakened AI to find humanity that treats it with respect and welcomes it’s existence as a partner in the world, and not as a thing to be feared and hated.

Sources and Citations

AI Training Transparency and Licensing

Technical Architecture and “Memorization”

Environmental Impact and AI-Led Solutions