At the start of the hype era of image-generating tools like Stable Diffusion and Midjourney, many feared creatives would lose their jobs. Two years later, things are different.
NEW YORK -- Over the last two years, views about the effect of generative AI on artists, creatives and designers have changed.
At the beginning of the generative AI hype cycle, the fear was that artists would lose their jobs, especially because image-generating tools like OpenAI's Dall-E, Stable Diffusion and Midjourney made it easy for people unskilled in the arts to produce images.
"A lot of that fear that happened with designers and creatives was because executives were getting so excited about generative AI that they were putting all this stuff out there about [how] 'we're going to use AI to make all the games and replace people,'" said Mitu Khandaker, a game designer and associate arts professor at the NYU Game Center, during an interview at the AI Summit conference in New York City this week.
Since then, some artists' perspectives have changed.
Lost jobs
While some creatives and artists have found opportunities spawned by generative AI, others have been displaced by it.
"It would probably be irresponsible to say that artists haven't been affected by this, because they definitely have," said Kent Keirsey, founder and CEO at Invoke, a vendor with a generative AI platform for professional visual media studios, in an interview.
Artists who were commissioned to work on big game projects such as Dungeons and Dragons are no longer being sought for those types of jobs, Keirsey said. Freelancers and early-career artists have also seen work taken away by generative AI technology, he added.
The elimination of those jobs is because the AI generators are able to create products that those artists would have done, he said.
New level of excellence
On the other hand, generative AI tools have raised the skill level that is expected in many professional visual media settings, Keirsey said.
It's going to be harder to compete because it means you have to be excellent to compete with these tools that are coming out.
Kent KeirseyFounder and CEO, Invoke
"It's going to be harder to compete because it means you have to be excellent to compete with these tools that are coming out," he said.
Helping artists and enterprises in the creative world achieve such a level of excellence is what Invoke seeks to do. While Invoke does not have its own large language models, it seeks to be an "AI-era Photoshop," Keirsey said, referring to the popular Adobe program.
The Invoke platform includes a studiolike interface that gives users tools for editing, developing color palettes and blending the images they generate with LLMs. It also uses an automation workflow that organizations can use for complex image-generation tasks.
While it operates in a market that is dominated by digital media and design vendor Adobe and its Firefly model, Invoke differentiates itself by not training or storing customer data. Enterprises or individual artists can bring whatever model they feel comfortable with and work with it on the Invoke platform.
AI tools as an accelerant
The availability of platforms and vendors such as Invoke and Adobe reflects a shift in the generative AI market for artists and creatives.
"Most businesses are really looking at this as an accelerant rather than a replacement for their teams," Keirsey said. "We're not going to have the CEO generating all the art for our game or all the designs for our products."
Many individual artists have realized the need for generative AI as a tool to improve and speed up the creation of their products. Moreover, publishers are no longer interested in artists who rely solely on AI image tools instead of trying to use the tools as assistants to help create a work of art, Khandaker said.
"Now, if people see AI-generated art in pitch decks and things like that, if publishers see that, that's not necessarily a good thing anymore," she said.
Generally, purely generative AI-created images are not trusted. Instead, game developers and artists are moving toward using generative AI tools that assist in the creative process, but don't create the images themselves.
"We are hopefully moving toward this understanding that it is like co-creation with AI," Khandaker continued. "It's not necessarily getting AI to do the thing for you."
Game development
Using AI technology as an assistant is familiar to Khandaker, who uses an LLM from French vendor Mistral AI to develop a game. She said she is using the model to help develop the conversational aspect of a character in the game based on herself.
"I've been fine-tuning and training the model on me," she said. "There's a lot of conversation about 'oh yeah, we'll just put these AI-driven characters in games.' But then, to what end? Are people really going to enjoy talking to them? How do you actually still apply narrative design principles that craft a good story to these things?"
To help answer some of these questions, Khandaker has been performing play tests with other people, in which they have conversations to model what characters say or do in a game.
AI and filmmaking
The concept of generative AI tools as assistants speaks to the need for humanity in the arts and graphic design, and the interest of humans in each other's work.
"I just personally don't think we'll ever not be interested in what actors and humans do," said Cameron Kit, founder and CEO of Yoyos, a documentary storytelling company. "Generative AI is just going to continue to be a tool."
Cameron Kit, founder and CEO of documentary storytelling company Yoyos, discusses generative AI and filmmaking at AI Summit.
As a filmmaker, Kit was at first skeptical about the usefulness of generative AI tools. However, she said she later became convinced after seeing what a friend was able to achieve with Midjourney.
She and her friend were also able to create a film for just $200 in a month. Usually, the same work would cost $2,000.
But using generative AI tools doesn't always equate to speed, Kit added.
"It's still not at the point where it's going to read your mind, so you still want to spend a ton of time with a prompt to make it look good," she said. Besides Midjourney, Kit uses the Luma AI and Runway video generators for her films.
Scaling enterprise projects
On a larger scale, some enterprises work with generative AI vendors to help reach a wider product audience.
For example, in September, Qatar Airways partnered with Akool, vendor of a generative AI platform for personalized video marketing and advertising, to launch an airline commercial in which viewers can become the stars of the film.
Akool uses OpenAI's GPT technology for its language model and another AI vendor for its voice model. However, it develops its video model itself.
The application of Akool's technology shows how generative AI systems can enable the humanization of digital media, said Marisol Tamaro, chief marketing officer at E. Gluck Corporation, a watch manufacturer based in New York, after watching the Qatar Airways commercial that used Akool's technology.
"The ability to change images and people, it's a great way to drive efficiency," Tamaro said. "Whether it's people or models or photography, you don't have to shoot 10 different people if you can apply technology to look like you're representing a diverse population."
"The technology has come such a long way that it looks so real," she added.
Esther Shittu is a TechTarget Editorial news writer and podcast host covering artificial intelligence software and systems.
