A.I - Avoid or Adapt?
By Niamh Martin.
An AI image generator is an artificial intelligence program that turns lines of text or prompts into hyper-realistic graphics. These tools, such as Midjourney or DeepAI have been subject to scepticism from industry professionals and hobbyists alike since bursting onto the markets in recent years. Others are more optimistic about a future where AI unlocks time and cost efficiencies to the benefit of all, as well as lowering barriers of education, and expensive design tools.
AI is capable of creating a wide variety of images in a seemingly unlimited array of styles, surpassing traditional artists and creators in speed as well as accessibility. To create the art, you simply have to prompt a visual, no one is inhibited by skill level or physical tools.
It’s a common opinion that every piece of art, or creative, is somewhat iterative or inspired by another piece. Still, with the introduction of AI Art the line between “inspiration” and “plagiarism” is becoming increasingly blurred.
Human vs. The Machine
One of the most common criticisms of AI Creative is the art typically generated is created with the use of unconsenting artists’ images. To create any image, AI leans into its database of art, full of artists who have no idea they’re included. In this case, there’s no real author, as the individual generating the art is using hundreds of artists to create any one image. In some cases, those who curate the databases behind the generator claim to "use only licensed images". Those with permissions, such as OpenAI, which claims its DALL-E 2 is trained on around 650m images, take from a mix of publicly available sources and these licensed sources. However, the company has not made the dataset public, leading to concerns that copyrighted material could be within it. This is important to note, as other machines have come to blows with reveals that hidden data used to train them, may be copyright protected.
The discussion around licenses, and copyright claims, suggests that the issue with AI is that neither the generator nor “prompt engineers”, have any real claim to being the creator of the art. In August of 2022, an AI-generated image won first place in a county fair, beating out other traditional artists to the cash prize. Outrage online ensued, when people discovered the entrant did not clarify in his painting description that it had been created through AI. Leading to widespread discussions surrounding the copyright disputes of AI-generated art.
More recently, on the 8th of November 2022, the U.S Copyright Office revoked protection to Kris Kahtoanova’s published comic book, Zarya of the Dawn in September, with the creative compiling solely of AI-generated art. The USPTO stated that copyrightable work requires human authorship. This could present problems within advertisements, were the tool to be used for a campaign.
With all the copyright issues considered, it’s important to highlight that when it comes to taking inspiration, individuals have a personal responsibility to honour the work that they be potentially emulating, whereas machines do not. The impact that labour or work has on a finished creation cannot be disregarded, be it in copy, art, or any field that machine learning can not emulate. By replacing the authors and creators of work with a machine that does not create necessarily better results, we instead could be putting all artistic work at risk of becoming derivative. Only with physical work can we come up with new solutions and ways of working, at the core of creativity is for people to view and edit their works, losing this would impact the range of artistry out in the world. There is no replacing the multi-faceted work a person can deliver.
“AI-created art is plagiarism imo (in my opinion). computers cant draw as I understand it, they're basically photobashing. People draw from memory/and impression of what they see. computers don't get impressions, they straight up use the thing. they have perfect memory and can’t "reference"”. - @alewph, Twitter.
Photobashing is a technique that consists of using multiple digital assets and textures to create realistic-looking artwork. Often resources needed for the process require licensing, or a creator may choose to limit themselves to copyright-free assets to use in their process. A lot of artists opt to use photobashing to model out a scene they will later paint, or illustrate, and the tool itself still requires a level of skill to navigate and complete.
The comparisons between Photobashing and AI are easy to make, they both generate content from other people's creative work, however, AI takes further steps in order to make “creative” standalone work by itself, versus photobashing creating assets to be used in conjunction with orders to create a finished piece.
Reinvention or Revolution?
“As a traditional oil painter I've always thought that this revolution was just around the corner. Ready to disrupt the photo and digital art spaces the same way photography disrupted traditional oil painting in 1880”. @ChambersFineArt, Twitter
Creatives have always evolved with new techniques and technologies in order to allow for innovation and new forms of creative expression. We can see from the adoption of photobashing in digital art, that this can be accomplished. Therefore, the conversation needs to move in a direction whereby we adopt AI in art as a tool, and how we go about using it in the creative process, rather than being the end product, and how laws can be produced to protect artistic property in this case.
The majority of artists probably won't work with AI at all, or quite rarely. Those who opt to become "pure" AI artists to generate ever and ever more realistic images out of AI will be limited to the confinements of copyright infringement, as well as the ethical questions around the tools. Even those who adapt to use it as a tool within their process will be dancing around this until AI laws progress further.
The Thinkhouse Take
“In 2023 we will see mainstream adoption of AI. AI represents one of the biggest opportunities we face to transform how we do things- from copy, to content to visuals. The ethical considerations of AI will continue to be widely debated and ultimately we must adopt an approach to AI that is respectful of artists and craft”. Jane McDaid, founder and CEO of Thinkhouse
“Another element of the AI conversation is from an influencer perspective. Virtual influencers. Some of them are obviously not real people they’re characters or cartoons, others look like real people. Both are taking jobs from actual people. Another element that is sinister about it is that white people can create a racially ethic virtual influencer and make money off their inclusion in a campaign that looks like it’s being diverse. When someone from that race/culture etc has not benefited in any way. It’s quite deceiving and unethical”. Donna Parsons, PR & Advocacy Director, Thinkhouse
“Depending on the tools used, your AI may either be trained on an untraceable endless volumes of art, or working off direct references provided. In either case, it's scraping the visual data, the visual creativity of these source artists, and spitting out something that will never benefit or acknowledge the original creators of the references”. Creative Designer, Thinkhouse
(Hotpot.ai's take on the Thinkhouse Disco Ball).
The AI Take
“As with many technology advancements, the answer to whether AI should be avoided or embraced in the marketing industry depends on a number of factors and can vary depending on context and perspective.
On one hand, AI has the potential to greatly improve many aspects of marketing, such as by enabling more effective and personalized targeting, streamlining processes, and providing valuable insights and data-driven recommendations. By automating routine tasks, it can also free up marketers to focus on more creative and strategic initiatives.
On the other hand, AI also raises important ethical and societal questions, such as the potential for biased algorithms, the need for transparency in decision-making, and the risk of job displacement. It is therefore important for the marketing industry to thoughtfully consider the benefits and potential downsides of AI, and to carefully implement it in a way that supports both business goals and ethical principles.
In conclusion, AI should not be avoided in the marketing industry, but rather embraced with caution and care. The key is to understand its capabilities and limitations, and to use it in a responsible and ethical manner that supports both business and societal goals.” - Chat GPT
AI generating work spreads further than just art or influencers, with copywriters, scriptwriters, or academics now becoming able to adopt it for writing capabilities as well. For example in the above paragraph, we enlisted the help of Chat GBT to form our opinions, which notably gave us a more positive-leaning opinion on the matter.
The outlook at the moment would point towards proceeding with extreme caution with the tool until clearer laws and usage rights are defined within it, however, these advancements in technology could be the beginning of a new normal within creative work.