Motifsnap

Swim on to the new AI creative landscape

AI art systems are popular. Despite the fact that they’ve been around for a while, 2022 will be recognized as the year the AI art revolution started. AI tech firms, large and small, for-profit and non-profit, have been building text-to-image generative models that have sent shockwaves through the creative community, which had previously felt protected from AI.

Even though the finest models are open-source (Stable Diffusion is widely regarded to be state-of-the-art presently), most consumers would happily pay a premium for no-code ready-to-use services. These premium applications will soon become widespread, and anybody working in the visual creative area will have to choose whether to learn/pay to utilize them or risk becoming obsolete.

Digital AI-savvy artists who understand how to use these models to expand their toolkits and better their work are ecstatic. A new world full with possibilities awaits them. However, some artists (who seem to be more conventional, but the borders aren’t clearly defined) disagree.

That’s what 3D artist and filmmaker David OReilly said on Instagram on July 22nd, when OpenAI announced DALLE would become a paid service:

Image 16

OpenAI scraped the web without compensating artists in order to feed an AI model, which would later become a service that those same artists would have to pay for. “It’s a shitdeal,” he adds.

Karla Ortiz, an award-winning artist, supports OReilly. On August 1st, she argued in a Twitter thread that firms like Midjourney — to which she specifically references — should offer artists the choice to “opt-out” of being expressly utilized in prompts designed to “copy [their] work.”

She ended the conversation with a cooperative tone that invites a much-needed debate: “I believe it’s an exciting moment, but this new technology must be managed ethically, honestly, and thoughtfully!”

RJ Palmer, a concept designer and illustrator, expressed his worries on Twitter as well. He just posted a brief thread on Stable Diffusion. “What distinguishes this AI is that it has been specifically trained on contemporary professional artists. “It even attempted to imitate the artist’s emblem that it copied from,” he claimed. “I am tremendously worried as an artist.”

Are OReilly, Ortiz, and Palmer just conservative artists who resist and reject technological progress? No, I don’t believe so. The issue is even more complicated. Palmer’s initial Tweet received 85K likes in a single day, which gives you an idea of the scale of their worries — and the degree to which other people agree with them.

That is not to suggest they are always correct in their arguments, but they are correct in that there is something to dispute about. We can’t go ahead until we have this dialogue if we want to establish a welcome future for everybody.

The road to AI generation

It is unclear what we should do with this technology. Should it be governed? Should anybody have free access to it? Should firms compensate artists whose work was used to teach the models if it is paid? Should individuals be able to imitate the styles of artists with only a few words? There are just too many unsolved issues for businesses, users, and regulators to ignore.

Artists, illustrators, and designers face increased competition from persons who may freely replicate or imitate styles due to a lack of appropriate legislation. For the same reason, AI businesses may train, deploy, and sell these AI art models as paid services.

AI art models that are opaque, stochastic, easily copied, and injudgeable cannot be subjected to present conceptual frameworks. They belong to a new kind of tool. The unique nature of AI art models, along with the absence of regulation, creates an explosive combination that makes this scenario particularly difficult. And, as Ortiz contends, it is critical to speak about, debate, and discuss until we reach a consensus that generates common ground.

That’s how we’ve always approached other technology. Hopefully, AI art models will not be any different. But they move quickly, and regulatory organisms must keep up if they are to catch up before the pandemonium consumes us all.

Shopping cart close