As technology continues to advance, there are increasing concerns about the impact of Artificial Intelligence (AI) on the art world.
With the ability to generate images from text prompts some worry that AI image generators could replace human artists and profit from their work.
There is also debate about whether the output of these AI systems should be considered “art” and whether artists should be compensated for their work being used to train the AI.
Additionally, there are concerns that the use of AI in the art could perpetuate harmful stereotypes and lead to the creation of problematic images.
Despite these concerns, some argue that AI can enhance human creativity and produce unique, original images.
“Revolutionary changes in the art world are not uncommon, but some believe this latest shift may be irreversible,” said Jason M Allen, winner of the “emerging digital artists” category at the Colorado State Art Fair, in an interview with the New York Times.
Allen’s winning piece, “Théâtre D’opéra Spatial,” was created using Midjourney, an AI system that generates images based on text prompts such as “an astronaut riding a horse.” Allen added, “Art is dead, dude.”
Many artists were outraged by Jason M Allen’s statement to the New York Times that “A.I. won. Humans lost” in the art world.
Despite earning only $300 from the Colorado State Art Fair, Allen’s words hit a raw nerve among artists who were already concerned that AI image generators could replace them and profit from their years of honing their craft.
California-based movie and game concept artist RJ Palmer tweeted, “This thing wants our jobs, it’s actively anti-artist,” a post that was liked over 25,000 times.
In other tweets, Palmer highlighted how well AI systems could mimic living artists, even attempting to replicate their signatures.
The AI image generators are indeed impressive, but they are based on the work of human creators.
These systems are trained on millions of images created by humans. For example, Stable Diffusion, a recently launched open-source AI image generator, learns from a compressed file of “100,000 gigabytes of images” sourced from the internet, according to its founder Emad Mostaque.
Mostaque, a computer scientist with a background in technology and finance, describes Stable Diffusion as a “generative search engine.”
Unlike a traditional search engine like Google, which shows you existing images, Stable Diffusion can generate images based on anything you can imagine. He says, “This is a Star Trek Holodeck moment.”
AI Art in a Blink
Artists have always learned from and been inspired by others, but some believe that AI is not simply finding inspiration in the work of other artists but “directly stealing their essence,” as RJ Palmer, a movie and game concept artist, says.
He also points out that AI can reproduce a style in seconds, whereas it might take a human artist a week to replicate it.
However, Emad Mostaque, the founder of Stable Diffusion, an open-source AI image generator, argues that the technology is a tool like Excel and it will not put artists out of work.
He suggests that young artists should find opportunities in this growing sector and make money from it, as thousands of artists already use AI to inspire and generate income.
OpenAI’s DALL-E AI system, for example, is used by more than 3,000 artists from over 118 countries.
AI has even been used to create graphic novels, with one author describing the technology as “a collaborator that can excite and surprise you in the creative process.”
However, legal challenges may prove difficult despite some artists’ anger towards using AI in art.
Professor Lionel Bently, Director of the Centre for Intellectual Property and Information Law at Cambridge University, states that in the UK, “it’s not an infringement of copyright, in general, to use the style of somebody else.”
According to Professor Lionel Bently, Director of the Centre for Intellectual Property and Information Law at Cambridge University, an artist would need to prove that the output of an AI system reproduces a significant part of their original creative expression in a specific piece of their art used to train the AI in order to win a copyright infringement case.
However, only some artists would have the resources to pursue such legal battles.
The Design and Artists Copyright Society (DACS), which manages payments for the use of artists’ images, is concerned about the impact on artists’ livelihoods.
When asked if artists’ livelihoods were at stake, DACS Head of Policy Reema Selhi replied, “Absolutely, yes.”
DACS does not object to the use of AI in art, but Ms. Selhi wants artists whose work is used by AI image systems to generate income, to be compensated fairly and to have control over how their work is used.
“There are no safeguards for artists [..] to be able to identify works in databases that are being used and opt-out,” she adds.
Artists may be able to claim copyright infringement when an image is taken from the internet and used to train an AI, although legal experts have stated that there are several reasons why such claims may not be successful.
Ms. Selhi also pointed out that proposed changes to UK law would make it easier for AI companies to legally obtain artists’ work from the internet, which DACS opposes.
Mr. Mostaque acknowledged the concerns and frustrations of artists, noting that similar issues arose with the advent of photography.
He stated that Stable Diffusion is working with technology industry leaders to establish methods for artists to upload their portfolios and request that their styles not be used by online services utilizing this and similar technology.
Bias and deep fakes?
Google had previously created an AI called Imagen, which could generate images from text prompts, but it was never made available to the public due to potential risks of misuse.
Google had warned that the datasets of scraped images used to train AIs often included graphic material, reflected stereotypes and contained “derogatory, or otherwise harmful, associations to marginalized identity groups”.
Recent concerns have been that Stable Diffusion could be used to create non-consensual nudity, deepfakes, and other problematic images, as reported by Techcrunch.
Mr. Mostaque stated that such unethical use would “break the license terms” and that the software already filters out attempts to create inappropriate content, although it can be circumvented by those with technical expertise.
However, he emphasized that the responsibility falls on the person committing an illegal act and that other tools, such as Photoshop, can also be misused, for example, someone could use its merge tool to superimpose someone’s head on a nude image.
Goo or art?
Sci-fi artist Simon Stålenhag tweeted that AI art is “derivative, generated goo” that the “new tech lords” want to feed us.
However, some notable figures are involved in the development of this technology. Elon Musk, who is known as the “techno-king,” is a backer of OpenAI.
OpenAI argues that DALL-E, their AI art system, enhances human creativity and produces “unique, original images that have never existed before.”
For a different perspective, I reached out to contemporary artist and broadcaster Bob-and-Roberta-Smith, who has exhibited in major galleries and will be taking over a shop at the Tate Modern gallery in London in October.
He primarily works with traditional media, but AI could be an interesting area of artistic expression, similar to the concept of a mash-up.
However, he believes that policymakers need to establish fair rules to ensure that artists are not taken advantage of and that money is not solely funnelled into the pockets of large corporations.
The use of AI in the art world raises a number of ethical and legal issues. While some see it as a tool that can enhance human creativity, others worry that it could replace human artists and perpetuate harmful stereotypes.
It is important for policymakers to establish fair rules to ensure that artists are compensated for their work being used to train AI systems and that they have control over how their work is used.
Additionally, measures should be put in place to prevent the misuse of AI in the creation of problematic images.
As the technology continues to evolve, it will be important to have ongoing dialogue and collaboration between artists, technology experts and policymakers to address these challenges and find a balance between protecting artists’ rights and harnessing AI’s potential in the art world.