What Are the Human Implications of AI and Creativity?

NAB

Generative AI will iron out dversity, human error and human effort, leading to a disastrous homogenization of culture that devalues the content, claimed artists including Grammy Award-winning musician Alex Ebert and digital artist Don Allen Stevenson.

article here

In a Fearless Media podcast hosted by Peter Csathy of Creative Media, they attacked AI for leading to a morass of artistic mediocrity and conformity.

“I’m going to be that voice [which says] it’s going to diminish the quality of our artistic output,” said Ebert. “There’s a very strange inverted relationship between democratization of taste and homogenization of output.”

He decried the idea of artists “reduced to simply a [human] being that prompts” an AI to create.

Ebert is the lead singer and songwriter for the American bands Ima Robot and Edward Sharpe and the Magnetic Zeros. He also scores films and won a Golden Globe Award for composing the music to 2014’s All Is Lost.

Tweaking a movie with test screenings is a long standing tactic by studios to “correct” by recutting a film before release. Generative AI could radically amp up that process — to Ebert’s horror.

“It’s suddenly like every other movie you’ve ever seen because the process of democratization actually leads to homogenization. You end up with a median [average] opinion. And I’m afraid that that is just what’s going to happen.”

Don Allen Stevenson, a multidisciplinary digital creator and crypto artist, agreed and thought that cultural homogenization would send artisanal creations onto a higher artistic and financial level.

“If everyone is able to generate so-called ‘high quality,’ AI-driven art, video or music and if the cost of those things is only represented as digital assets, it will reduce the quality. But I think simultaneously it would increase the value of physical things that are more tangible.”

This is one theme of the metaverse bible “The Diamond Age” by Neal Stephenson. In his science-fiction there are 3D printers that can print anything on demand from a text prompt.

“So it made the cost of materials very low,” said Stevenson. “But then what people seem to value in this example was stuff that was handmade. They loved like that. The elites and the rich in this novel loved their handcrafted things because they were truly unique.”

“That’s why distortion became interesting,” said Stevenson referring to the idea that much of what we appreciate about art stemmed from a mistake.

“Constraints are how originality occurs. So where are the limitations of AI? And when we find the limitations of AI it might become interesting. The only cool things I’ve seen AI spit out so far [is when] the AI fails — where it can’t do fingers and it makes all these weird images and where ChatGPT is spitting out nonsense.

Ebert said he doesn’t use AI to produce music. “It’s honestly not that much fun. It’s quicker and more productive, but it’s not as interesting for me. I don’t reach interesting limitations. I don’t end up with an interesting sound that you could never recreate because of the reflections in the given room [or the way I’m playing a particular instrument on that day in that room]. “These constraints, these failures are important.”

So, where are the constraints and failures of AI that will be interesting enough to forge your own path apart from it?

He argued that humans still have an affinity for the idea of an object or piece of content with tangible origins.

“If we see something artisanal we’re like, ‘that was made by hand and it’s a one of a kind and it makes me feel special because it is special.’ But in order for that to happen, you have you have to have a sense of a tangible origin.”

Yet, we’re so beguiled by imitations of tangible origins? “We’ll buy the pre-ripped jeans, we will buy an experience of struggle. We’re buying the thing at Urban Outfitters that looks like it’s from Peru and made by hand but [in fact] it’s mass produced to look like it’s from Peru.”

Following on from this, the panel pondered whether proof of human craft in producing art would be required in order to validate its artisanal value in the age of AI.

Stevenson suggested an artist could document their process to “show what creative human made decisions were made, what was the intentionality, what was the heart?

“And if there were a legal structure that could look at that when judging the output, like how much human level work went into that generated thing.”

Stevenson added that he’s been encouraging people to live stream and document and record voice memos in the process of creation to act as a chain of proof.

“Have people interview you as you’re making whatever art form you’re making and then have that be a part of the art piece, have that be the story.”

He continued, “Humans are very narrative based. We love story. So, if you don’t have a story that shows that you put human level love, energy and heart into that thing, then you’re just an AI generated automation, homogenous nonsense. But if it’s like, Wow! this person put a lot of actual blood, sweat and tears into this and we can measure that, we can record that, then maybe [that might work].”

Csathy summed up, somewhat fatalistically, “I want to believe that no matter how sophisticated AI gets, there’s something about humanity that will be appreciated and that will be differentiated, so we convey the humanist aspects of it. We all just have to be very stoic about the fact that this is happening.”