In just a few years, artificial intelligence (AI) has become a significant development, gradually integrating into various fields.
One field that has experienced its impact and has raised concerns is the creative space, with some users embracing it while others remain doubtful and worried about its impact on their job security.
Wedding photographer Lennon Tan shared that fellow photographers have increasingly used tools like AI Photo Editor in Adobe Photoshop and Imagen.
Tan noted that the wider the availability of tools, including AI, the better it is for the industry. — LENNON TAN
“Imagen allows you to train the AI and have it edit your whole album according to your style.
“What would take us four full days of editing will now take us four hours, if done correctly,” he says.
Travel photographer Grey Chow echoed a similar sentiment, noting that AI-driven tools have been present in the industry for a long time.
In his perspective, AI generally falls into two categories: tools that generate standalone images and those that can be integrated into a photography workflow.
“The first (image generation) will be entirely different from photography; you can use AI to generate all sorts of images, such as a realistic portrait, a beautiful landscape, a logo or a cartoon sticker.
“The second way is using AI as part of our photography, with certain AI technology already existing for years, such as AI-driven autofocus (for cameras) or using AI to upscale the size of an image while retaining the details or streamlining the photo editing workflow,” he says.
Photography perspectives
In the case of AI-enhanced images, fine arts and film photographer Emilio Daniel says the debate is ongoing about their use in the industry, though he personally views them as just an additional set of tools.
According to Emilio, common uses of AI tools in photography involve automatic touch-ups and replacing, say, the sky in images. — EMILIO DANIEL
“Honestly, I don’t really have a problem with them, as they are simply post-processing tools that are automated to do things that a human would do anyway.
“They are just a set of instructions to be repeated; most of the time, people who use these AI tools will make their own tweaks after applying them,” he says.
According to Emilio, some of the most common applications of tools like these involve automatic touch-ups and replacing, say, the sky in images.
Chow expressed similar sentiments, pointing out the value of software tools like upscalers, subject and background removal, and noise reducers in enabling photographers to produce better work.
Despite these innovations, Tan says that people will still have to capture moments and direct subjects to pose, and there are many “human factors” that AI can’t replace.
However, at the center of this debated topic is the use of images entirely generated by AI, especially when passed off as authentic photographs.
To make a statement on this topic, a Berlin-based photo and video artist who had won the creative open category at last year’s Sony World Photography Awards organized by the World Photography Organization declined the prize.
Boris Eldagsen revealed that his winning piece, Pseudomnesia: The Electrician, was actually the work of AI, and he had participated in the competition as an experiment to spur discussion about the future of photography.
“How many of you knew or suspected that it was AI-generated? Something about this doesn’t feel right, does it?
“AI images and photography should not compete with each other in an award like this. They are different entities. AI is not photography. Therefore, I will not accept the award.
“I applied as a cheeky monkey to find out if the competitions were prepared for AI images to enter. They are not.
“We, the photo world, need an open discussion. A discussion about what we want to consider in photography and what not. Is the umbrella of photography large enough to invite AI images to enter – or would this be a mistake?” he says in a statement.
The World Photography Organization later asserted that it was aware the image was AI-generated, although this information was not disclosed until after Eldagsen’s speech at the award ceremony.
Eldagsen is not alone in wanting more discussion over the use of AI in the industry.
Those like Emillio want to have clearer distinctions between actual photography and what many in the community are calling “promptography”.
He says that AI-generated images, even with artistic merit, are fundamentally not photography.
Chow divides AI into two types: tools for creating standalone images and those aiding in photography workflows. — GREY CHOW
Chow, on the other hand, expresses concerns about the striking realism of such imagery, stating, “even as a photographer, I can’t always tell if it is a real photo”.
“My concern is that people may abuse it by providing false information and even posting it on social media for manipulation. Other than that, there’s a grey area about how AI-generated images should be used.
“Some stock photo market platforms only accept images generated by AI that do not use any well-known people’s faces and do not contain any artist’s name in the prompt to avoid copying someone’s style.
“Still, I think that’s not enough. We will need certain authorities to implement rules to avoid the abuse of AI,” he says.
For Emillio, the troubling aspect of AI is that it can be trained on artists’ work without permission, effectively “stealing from them”.
Art attack
For casual users, the generative AI tools can serve as delightful distractions, enhancing activities like making character portraits for Dungeons & Dragons and generating whimsical images.
For those whose livelihood depends on art commissions, the narrative takes on a different tone, with ethical concerns about how these tools were trained coming to play, adding a layer of complexity to the discussion.
A Malaysia-based concept artist and illustrator who wanted to be known only as Wong touched on the rise of copycats due in part to the easy accessibility of the AI generation.
With over 41,000 followers on Instagram, Wong undertakes art commissions and contract work from various industries.
“If AI as a tool can be used as support for actual artists – for example, to speed up certain processes – instead of a way to make complete art and sell it, I wouldn’t find it particularly concerning.
“I just dislike how the art they take to train AI models is usually stolen. I think most people who use AI art at the moment are not artists needing tools, but rather non-artists wanting quick and easy income,” she says.
For casual users, the generative AI tools can serve as delightful distractions, enhancing activities like making character portraits for Dungeons & Dragons and generating whimsical images. For those whose livelihood depends on art commissions, the narrative takes on a different tone, with ethical concerns about how these tools were trained coming to play, adding a layer of complexity to the discussion. — Image by vecstock on Freepik
AI image generators such as Midjourney and Stable Diffusion are trained on a dataset of artwork, with one of the most commonly used being the Large-scale Artificial Intelligence Open Network (Laion).
Laion 5B, the specific dataset used in AI generators like Midjourney and Stable Diffusion, comprises 5.85 billion images and has been the subject of debate in the creative community and even lawsuits.
This was the case for American artist and illustrator Kelly McKernan. Using the online tool “Have I Been Trained”, she uncovered that the Laion dataset included more than 50 of her artworks.
Following this, McKernan and a number of other artists filed a lawsuit against the generative AI businesses Stability AI, Midjourney AI, and Runway AI as well as DeviantArt, an online community and platform for artists to display their work.
The lawsuit (Andersen vs Stability AI) is ongoing, and one piece of evidence includes a list of artists that Midjourney allegedly used to mimic a specific style.
The extensive list comprised more than 4,700 names, including notable figures like manga artist Yusuke Murata, acclaimed for his work on One-Punch Man, and Banksy, the elusive England-based street artist and political activist.
It even included art legends such as Vincent Van Gogh, the Dutch post-impressionist painter celebrated for his vividly expressive works like Starry Night and Sunrise, and Andy Warhol, the iconic American pop artist renowned for creations like the Campbell’s Soup Cans.
Concerns surrounding the use of existing artwork in training AI models led to the creation of Glaze, a tool developed by the Sand (Security, Algorithms, Networking, and Data) Lab at the University of Chicago.
In essence, Glaze conceals the style of an artwork by computing a series of alterations, causing an AI model to perceive it as an entirely different art style, essentially “poisoning” the dataset.
Additionally, a research paper from the Stanford Internet Observatory, published in December, revealed that child sexual abuse material (CSAM) was included in the training dataset of Laion 5B and, consequently, in its derivatives.
Creatively co-existing
On the business side of things, comic artist and illustrator Lipwei Chang says that although AI is a concern within the community, he personally hasn’t sensed its impact.
“I believe it depends on which field and what kind of clients you have. In my case, I’m dealing with collectors or those who want exclusive art, so there’s more of a preference towards originality.
Chang says that breaking into the industry has always been challenging, with or without AI. — LIPWEI CHANG
“There may be some concerns right now, but AI is only at the beginning. As time goes on and we keep producing genuine art, AI will expose its own problems, such as a lack of human artistic touch,” he says.
Chang, who has over 13,000 followers on Instagram, says he views AI as another artist that is fast and able to produce pieces in large quantities.
However, as it has to be fed data and instructions by someone else, he feels that it is lacking the ability to experiment and make decisions, which results in a less natural final result.
On whether AI’s proliferation would make it harder for up-and-coming artists to enter the industry, Chang says that breaking into the industry has always been challenging, with or without AI.
He believes that not much has changed, emphasizing that newcomers still need to invest the same level of effort to establish themselves.
Wong similarly says that she has not noticed any particular effect on her online commissions.
“I think it’s the opposite; I get more clients, especially since a lot of clients now want extremely detailed and custom work that I think would be hard for AI to achieve,” she says, adding that some also view AI art as less desirable, preferring unique hand-drawn styles.
They both agree that there is a place in the creative world for AI, with Chang viewing it as a potential tool for education and for beginners in the initial stages of rough or draft work before they accumulate enough experience to start sketching independently.
Meanwhile, Wong recommends focusing on tools that can enhance creative workflows by handling mundane tasks such as sorting and renaming layers, aiding in the creation of in-between animations, and serving as temporary stand-ins for the final piece.
“AI can be a good tool, as long as it is not the only medium to achieve the final piece with zero additional creative or technical input.
“The world will always have new tools and new ways of creating art. The appearance of digital artists and tools like Photoshop did not replace the value of traditional artists.
“As long as you are proud of your craft, have a solid portfolio, and show professionalism, you can break into the industry you aspire to be in. AI is just a tool, it does not create new ideas,”she says. Her advice to new artists is not to fear entering the industry due to the potential of AI replacing them, even if AI-generated artwork may appear superior to their current skill level.
She stresses that the unique ideas newcomers bring are what hold value and are irreplaceable, and that technical art skills will develop with practice.
