While AI offers tremendous potential for the creative spheres, ranging from the art market to tech startups and everywhere in between, retaining a human touch is necessary – and that begins with human judgment.
Right now, in certain cases it’s easy to tell when an image has been generated by AI. Hands – already a challenge for human art students – continue to baffle the computer, which churns out digits that look like, as one journalist writing for The New Yorker put it, “diagrams in a medical textbook from an alien world.” Similarly, faces in AI drawings can take on a haunting ghastliness like something akin to The Scream: smudged, skin-colored forms with dark crevasses for orifices. Visual artists, inspired by AI’s unique aesthetic, have integrated artificial images into their oeuvre, prompting some controversy when such works receive prizes.
In other instances, such as an allegedly AI-generated song posturing as a collaboration between musicians Drake and The Weeknd, industry members are beginning to worry that creative professionals may soon be outstripped, or even replaced, by the machine. Is it possible that AI creativity could supplant human creativity? More broadly, how will AI aesthetics affect art?
Many working within the AI industry, like David Holz, the founder and CEO of Midjourney, have gone on the record clarifying that they do not see AI as a substantive threat to creativity and artmaking. “The goal is to make humans more imaginative, not make imaginative machines,” Holz told Forbes. “I think [that’s] an important distinction.” Similarly, Microsoft CTO Kevin Scott sees AI as a “copilot for everything,” a tool that can help – and not replace – creative-driven people.
The dilemma of using AI tools for creative expression
Imagine you’re looking over your bills and need to do some math. Rather than pulling out pen and paper and performing time-intensive equations to reach your sums – leaving ample room for error – many of us prefer to tap numbers into a calculator, which does the same work perfectly, in a fraction of the time.
For AI professionals like Microsoft’s Kevin Scott, tools like ChatGPT offer creative minds a similar stepladder to realising their artistic visions. As he put it,
I’ve been playing around with an experimental system I built for myself using GPT-3
designed to help me write a science fiction book, which is something that I’ve wanted to do since I was a teenager. I have notebooks full of synopses I’ve created for theoretical books, describing what the books are about and the universes where they take place. With this experimental tool, I have been able to get the logjam broken. When I wrote a book the old-fashioned way, if I got 2,000 words out of a day, I’d feel really good about myself. With this tool, I’ve had days where I can write 6,000 words in a day, which for me feels like a lot. It feels like a qualitatively more energising process than what I was doing before.
Many of us might feel our stomachs turning. Dividing sums using a calculator is one thing, you might think, but should the arduous work of producing art be outsourced to text-predictive robots, whose job – as Ian Leslie puts it – “is [to be] by definition generic”? If the world becomes filled with books peppered with generative, generic phrases plumbed from the depths of the internet, will our own creative imaginations become stultified in the process?
Some, like the writer and editor John Warner, offer the provocative suggestion that humanoid writing is already formulaic and mindless enough: when schools teach students to write simple essays in a rigid five-paragraph structure, he argues, “students no longer learn how to think through the problem of structure.” GPT3’s writing, complete with paragraphs, punctuation, and even articulate-looking sentences, is not unlike that of a lazy student: it’s learned perfect form, but features lackluster (as Warner puts it, “BS“) content.
It’s unclear how AI itself will affect human creative thought. In some cases, humans are modifying to meet the machine: pop songs are often created in concert with apps like Spotify, which use machine learning to identify traits of popular songs among users, and then push similar songs based on those qualities.
How AI will impact the process of creation
The image of an artist toiling in their studio – the piles of paper fluttering to the floor around them; ink-stained fingers and discarded canvases – is challenged by the existence of AI. Many of us would bristle against the idea that artmaking should be streamlined or made more efficient; that there is value to the artistic process, however long it takes.
In truth, reports of the death of human artmaking may be greatly exaggerated. A recent study conducted at Oxford found that most artists report “essentially unchanged” relationships to their arts, “as artists ultimately work to address human – rather than technical – questions.” Rather than replacing their process entirely, the researchers found, artists have been using machine learning for five main activities in their practice:
- Technical research
- Using and building machine learning models
- Using and building datasets, training models
- Combining models
- Curating outputs
Where close interaction with AI may lead us
Recent innovations in artificial intelligence have prompted daunting questions about the human, and creative, experience. If we continue to engage with AI in recursive loops, how will our relationships to ourselves evolve (or dissolve)? There is a chicken-or-egg process in the making: could AI, that – as Ian Leslie wrote – “is by definition generic,” actually reshape the way we perceive and communicate about the world? While the aesthetic influence of AI remains to be seen, experts are arguing that the artistic process – the imaginative expression of emotion and experience – will always be as human as ever.
Doron Fagelson is Senior Vice President, Media & Entertainment at DataArt.