Weyburn artists on generative AI: Tool or trouble?

Over the past few years, generative AI has become increasingly popular. From services like ChatGPT and Google Gemini to Adobe Firefly, to Suno, AI has been used to create everything from life-like images to poems to essays to songs. The usage of AI is resulting in questions being asked as well, particularly over some of the ethics of using it. 

Criticisms of using generative AI to create things range from taking away work from writers, artists and musicians to the works created lacking the soul that a person can pour into something. The idea of AI having an impact on the workforce was part of the reason for the months-long Writers’ Guild of America strike last year, as screenwriters were concerned movie studios would use generative AI to replace them.  

Another aspect of AI has been its use in the creation of deepfakes – images of public figures (and sometimes not public figures) that are 100 percent fabricated but look real enough it can cause someone to suspend disbelief.  

Some of the positive reception of AI is from those who see it as not the final creator of the works, but as a tool, whether through helping to get past a writer’s block, or putting a melody to lyrics of a song to help provide direction. Some also use generative AI to create images for things like children’s books or other media due to a lack of funds to hire an artist.  

The question of whether AI is something positive or not is not easily answered, either, with both sides of the argument present in the artistic communities.  

Dan Cugnet is a singer-songwriter from Weyburn. He approaches the use of AI in music as part of the evolution of the music process. He noted that in the 1970s when synthesizers were first coming into usage, the music created with the technology was considered by some to be soulless because it was digital, and not created with an instrument like the piano. 

“When a piano came after a harpsichord, was that soulless because it didn’t have the same format and the same strings?” asked Cugnet, as he reflected on the changes in instruments over the years. “I think a lot of the times, the magic in a song is accidental, even with an artist, you know? Sometimes it’s really constructed and construed in its process, but sometimes that magic or the secret sauce is just something that’s completely random and accidental in a lyric or a sound.” 

Cugnet touched on the creation of the song Jive Talkin’, a hit for the Bee Gees in 1975. Barry Gibb from the band said the rhythm for the song came from the sound the car made while driving across the Julia Tuttle Causeway in Florida. The song was also one of the first uses of what is called synth bass in a pop song. 

“That’s so accidental and so random, so was there soul in that car, and how it was contacting the asphalt, driving from his house to the studio? You just heard something, and it triggered something. You listen to the start of it, and it’s such a different, unique thing and it’s so iconic.” 

Much like how the sounds of the car helped to trigger the creative thought process that created one of the most popular songs of the disco era, AI can also help to trigger those processes in many ways, Cugnet said.  

“I think the AI is amazing for creating art, creating prose, creating music,” he continued. “Is it imperfect? Absolutely. Is it still being directed in an extension of us? For sure it is. No different than... Think of an electric guitar and a bank of pedals, and you’re plugged into an amp. Is that soulless? No. It’s still just an extension of what we are, and I believe that with computers, computer software, it’s the same thing, right?” 

The technology is a key part of the evolution of music, Cugnet highlighted. The technology has adapted to where people can have access to different apps right on their phones that can be used to make music, and generative AI is just another one of the tools that can be used.  

“I love AI and I love technology, and I think that it’s, to me, a way of... I don’t want to say it’s levelling the playing field, but I think it just creates openings and opportunities for people that normally wouldn’t, and that’s the same thing.” 

*****

What are some of the possibilities of generative AI? I used the image generator Adobe Firefly, the music generator Suno, and Google Gemini to create some examples. 

Adobe Firefly generated this image with prompt: "realistic, a city street at night, far perspective, skyscrapers in the background, neon lights reflecting off puddles, two people walking away from camera with umbrellas"

Adobe Firefly created this image from the prompt: "Photo of a baby in a sweatsuit, bench pressing a lot of weight"

 

The lyrics for both of these songs were written by myself. The lyrics were then put into Suno, and the music, and the vocals, were created entirely by generative AI. 

So, what does a generative AI model think are the benefits of using it during the creative process? I asked Google Gemini with the prompt ‘Write 1-2 paragraphs about generative AI and the benefits of using it for the creative process.’ 

The response it gave illustrated some of the same benefits provided by Cugnet. 

“Generative AI can act as a powerful brainstorming partner, suggesting unique concepts and variations that might not have occurred to the creator,” the machine stated. “It's great for overcoming creative blocks. These tools can automate tedious or repetitive tasks (like creating different colour palettes or variations on a design). This frees up time for focusing on the bigger picture and strategic aspects of the creative process.” 

However, generative AI also recognizes its own limitations and criticisms. 

“Over-reliance on generative AI might lead to similar-looking and unoriginal content, the AI continued. “AI models inherently reflect the data they were trained on, which could lead to homogenization of styles. AI-generated creations can blur the lines of originality and raise questions about authorship. While the human provides the prompts, the true 'artistry' of work generated largely by a machine is debatable.” 

***** 

“It can actually do a lot of research for you, which is very handy,” said Leena Honig. She is an author and an artist in Weyburn. She explained how it can be used as a tool to create a reference image, similar to how one will take a photograph to use as a reference for painting a landscape scene.  

“If you want a rainbow and a mountain, and some deer or something, but you don’t know how to put it together, you can just type all that in there,” Honig said of using an application that can generate images. Just using those applications to generate images, in some contexts, can have negative impacts, however.  

“A lot of people publish children’s books, or just want content for stuff, but artists won’t get paid, right?” Honig posited. “Because it’s free. The app is free and it definitely could replace a lot of things.” 

When using AI as a creative tool, Honig said it does need to be disclosed that it is an image, or other work, generated by AI. 

The prevalence of generative AI for content generation on social media was something else Honig noted. 

“I think it will be a problem if you use it too much,” she continued. She pointed to the use of it on social media as something where it is starting to become a hindrance to the actual process. “You could tell that it’s fake because, like for interior design, things will be in odd places they’re not supposed to be, and it’s overtaking so it gets annoying.” 

Honig also pointed out one of the biggest flaws with generative AI – it isn’t coming up with the images and words it creates organically. It has been trained to associate words and images with other words and images. No matter what the prompt used is, it is still looking for things on the internet.  

“You could pay an artist to do exactly what you’re looking for, and then you have that artistic viewpoint,” Honig described. “Otherwise, it just kind of generates off of images that it’s seen through history, so it uses references itself.” 

Like Cugnet, Honig said it is a tool and should be used accordingly.  

“You wouldn’t be able to take that image and print it off on a canvas and call it artwork because that’s a computer program. It’s fun for crafting or you want to play around with photos at home or something like that. You know it’s a fun tool to just play around with, but it’s not really to be taken very seriously.” 

I asked Google Gemini to summarize things nicely and succinctly: 

Generative AI, like any technological tool, carries a unique duality. It holds the potential to streamline, inspire, and expand the boundaries of creation. Yet, like Cugnet points out with his electric guitar analogy, it remains an extension of the human creator, not a replacement. As Honig highlights, AI-generated work can lack the nuance, intentionality, and originality of work born from the depth of a human artist's experience. 

The very AI we've asked about its role in the creative process seems to agree. Its ability to brainstorm, automate, and even research makes it undeniably valuable. But it also echoes the concerns of homogenization and the blurring lines of authorship. Perhaps true art, no matter how technology reshapes its creation, will always require that 'secret sauce' that Cugnet describes – the accidental spark of inspiration only the human soul can provide. 

Ultimately, the question of whether AI is a net positive or negative in the realm of creativity may never have a definitive answer. Its power, like all power, rests in the hands of its users. As artists, musicians, and writers grapple with these evolving tools, it becomes vital to ensure that generative AI serves human expression, rather than replacing it.