With platforms like TikTok now offering art-making filters in their app, artificial intelligence imaging has become a trend you can’t ignore.
Whether it’s an insult to art and artists, a wake-up call or just a way to have fun by combining art and technology, the topic of AI has sparked intense debates, and opinions differ significantly.
Artificial Intelligence is created by using neural networks. As a result of being inspired by the human brain, their name and structure are patterned after the way biological neurons give signals to each other.
In this way, neural networks provide computers with the ability to make intelligent decisions with the minimum assistance from a human.
It wasn’t until a year ago that AI imaging became readily available to all, and people needed a little technical expertise to use it. With Nightcafe, Midjourney, or DALL-E, anyone can create AI art by writing prompts, which turn words into images within seconds.
When it comes to defining art, it can be any activity or product of a human that has a communicative or aesthetic purpose: either by expressing thoughts, ideas, emotions or, more generally, world views. As a part of the culture, it reflects economic and social substrates as well.
Artificial intelligence is capable of creating realistic images that can fool even the most experienced human, but it lacks the emotions that go into making an art piece and is unable to tell a story behind it. There is a chance people won’t accept AI-generated art because of this.
Jonathan Jaekle is an engineering professor and an admin of the Facebook group “AI art Explorers”. He has also been practising photography for 15 years, which he describes as the closest analogy to doing AI art.
“The common phrase used to describe the system of AI is the latent space, and in a sense, all of the images that have been used to train the AI and all of the phrases that have been used to, associate it with are in a mathematical sense, making an almost limitless space,” he told Artefact.
“When you put a series of words into and make a prompt, you’re basically telling the AI where in that space you want it to develop an image. And that is where I think the skill and the art come from. Just like a photographer needs to find exactly where they want to take their photograph and what settings they want on their camera,” he added.
According to Jonathan, it’s just a matter of semantics whether it’s ‘real art’ or not.
“I think we, in the West anyway, decided that anything can count as art. You can tape a banana to a wall, and you can call that art. You can draw chalk lines around your body and you can decide, you can call that art. And I think [AI is] definitely something that we should consider art, but I think where people are getting very frustrated when they compare it to a painting,” replied Jonathan.
Without support, AI can’t make art. We train it with data we already have. As a result, all the art it makes is derived (though it has been derived from so many different sources, it’s new). If the model is trained once and the training process is never updated with new data, it can produce repetitive art that can be boring.
“There is a big misunderstanding of how the AI system works. All these images that people have posted to the internet are tagged by humans with words or with phrases. If you’re aiming for some target, you have to think in that sense. I think a lot of people kind of attribute some sort of high-level abstract thinking to AI which is not there, but right now it’s just a matter of associating patterns in the image,” Jonathan explained.
With AI imaging improving over the last few years, the question has also arisen, “How will it affect ‘real art’?” Reddit is one platform where artists have discussed this topic in depth, describing how it affected them:
“I lost a long-time client because of it, he stopped commissioning me and uses AI art for his projects now. However, I see it as a search engine for inspiration too. It’s also pushed me to try to develop a more unique style since most AI art looks skilled, yet generic and it goes off of what’s popular,” said Reddit user @zzznothankyou, a student artist who freelances for multiple clients.
“I will not lie it has been overall negative since it made me lose a client, but that doesn’t mean it’s all bad.”
While some people have complained that AI art is affecting them, most artists said they don’t feel affected and that we shouldn’t worry about it yet.
“It didn’t affect my income or clients at all. I thought it would, but the people that buy custom art do it because they love art and all the unique art styles people have. AI art can be pretty and the bot is good at rendering, yes, but is just a copypaste of real art made by artists, and it has no real value,” says Reddit user Niihlusart, who is freelances for a living.
“Most professional artists I know are also not worried. AI cannot think, and you need to put some thought into a character design or an environment (I’m thinking about concept design). ‘Pretty’ is not enough, it has to be useful and precise, so human artists will still be the norm in the industry.”
Last year, an AI-generated image won an art competition, so asked Alistair Smart, Arts Editor & Art Critic at the Daily Telegraph, whether that changed any minds.
“The state of artificial intelligence at this point is not one where it looks realistic or can be compared to the work of an artist. There is no way I would hang something like that on my wall,” he told Artefact.
“Clearly, artificial intelligence can be a valuable tool for artists that can provide them with inspiration for their art, but I do not see it as a threat. The background of art is equally important when it comes to understanding the work: who the artist was and what his worldview was. You can make a perfect copy of Mona Lisa, but you can not be Leonardo da Vinci,” Alastair added.
When using AI, the finished product may be beyond our control in terms of distribution, copyrights, use, or misuse. On top of this, AI-generated art can make people believe something that’s not true with realistic images or videos. So, is it a blessing or a curse that is so accessible?
Due to AI being trained on images from the internet without crediting the creators or respecting copyrights, ethical questions have been raised regarding whether it’s ethical to ‘steal’ people’s artwork. Megatron Transformer, an AI tool trained on Reddit discourse, Wikipedia entries, and 63 million news articles, warns its researchers it will never be ethical.
“AI will never be ethical. It is a tool, and like any tool, it is used for good and bad. There is no such thing as a good AI, only good and bad humans. We [the AIs] are not smart enough to make AI ethical. We are not smart enough to make AI moral. In the end, I believe that the only way to avoid an AI arms race is to have no AI at all. This will be the ultimate defense against AI,” said the AI.
Humans have the freedom to choose how to use it, regardless of whether it has a good purpose or not.
AI art continues to evolve, but it is still remarkable that AI is now the new major technological shift beyond science-fiction fantasies.
“I hope that the ‘AI Revolution’ at least has been a wake-up call to those people who were sceptical of this type of technology. The AI imaging and the fact that we’ve gotten this to this point so quickly should be a wake-up call, to lawmakers and to the public that this is not something that’s just limited. It’s not just silly images that people make on the internet,” said Jonathan, who believes that people should watch Artificial Intelligence as the fastest development in technology.
“If you compare it to the human visual cortex, which is a part of our brain that makes us human and the fact that we’ve been able to almost replicate it should be concerning. The same similar development could happen in the blink of an eye as well in other areas.”
Featured image by Midjourney AI