Although most people would consider art as something that can only be appreciated visually or something usually found in museums, the truth is that we spend our daily lives interacting with art in some form or another. Whether it is from the way we dress to the cities we walk in, or in the way we express ourselves, we often tap into tools that are products of someone’s idea and creativity. From stone masons, and artisans, to modern technicians and visual designers, art and creative professionals are responsible for creating most items that people use and interact with on a daily.
Machines and innovation evolved from the simple need of people to solve their typical problems. Man’s endless pursuit of invention, creativity, and autonomy can be observed from the works of Da Vinci up to today’s development of artificial intelligence. Nowadays, machines have gone beyond solving simple tasks for people and have moved towards automation in production and content generation. Digital artworks, marketing collaterals, and other digital assets can now be easily produced in minutes with the help of AI-based platforms.
Despite having huge benefits for the digital marketing industry and providing a cost-effective solution for businesses, automation and machine learning still poses a threat to existing creative industries such as content writing and graphic design. To further understand how this innovation affects concerned segments of the industry, we venture out to dissect and look at the workflows involved in AI content production as well as the value it can provide to consumers:
A companion for writers?
Publishing used to be dependent on a hierarchy where writers and editors were the protagonists in delivering succinct and persuasive articles. Whether writing about their experiences or from expressing opinions on the works of others, ink and paper are the typical weapons of a scribe or a writer. Nowadays, writing are done with digital ink via computers and is no longer just a cerebral and creative hobby for a chosen few but also an indispensable tool in areas where the ability to persuade and influence others is much desired. No longer do people need a scribe to etch out their feelings and express their grievances, they can easily send a message out through their phone or email all while lining up for a cup of latte.
Poised as a tool dedicated to maximize marketing efforts, modern digital writing is more focused on hitting keywords and optimizing articles to reach certain demographics and cultures. With the introduction of ChatGPT, anyone now with an access to a computer and the internet is granted the capacity to concoct articles simply by churning out a few phrases or placing a few prompts with their desired keywords. For people in a rush or someone experiencing a writer’s block, having an instant online AI helper like_ ChatGPT_ must feel like an easy way to get out of a rut and get your work moving. But to understand its full potential and relative drawbacks, the best way is to put some ideas to work and compare how accurate AI can carry out a writer’s vision.
Experiment #1
Prompt: “Write Dr. Seuss’s Oh the Places You’ll Go in the style of Edgar Allan Poe”
Expectations and Observations
The overall output captures the gothic style of Edgar Allan Poe’s writing but failed to incorporate the storyline of Dr. Seuss’s work. The ideal would be to strike a balance between the original and the additive tone where Dr. Seuss’s rhyme and rhythm blends harmoniously with Poe’s gothic writing style. For example, the phrases, “Oh the Places You’ll Go” can be given with a bit of sinister twist such as, “Oh the Places Where You’ll Perish” and so on and so forth.
Here we can identify upfront how ChatGPT interpret prompts in a more literal translation or rather a transliteration of the work that results to a less meaningful final output. But of course, we might be expecting too much from a simple prompt or just haven’t done enough iterations to come up with the ideal tone and output.
Can it help reinvent how artists work?
In an already saturated market, AI-generated images pose a gargantuan threat to the works of traditional graphic designers and illustrators. Harnessing the power of intuitive text-to-image algorithm, AI-generated content is the product of machine-based learning tools that largely depends on existing data from online databases. With a text prompt as a starting point, online tools such as StabilityAI, Midjourney, and NightCafeAI can easily generate images online within minutes.
From a traditional point of view, using machine learning to create a product based on someone else’s existing work diminishes the sanctity and purpose of having art. It is like taking a piece of Picasso’s work, cutting a snippet from Van Gogh’s canvas, and piecing it together in Gustav Klimt’s style without having definitive reasons. As for any artist, getting your work copied without credit is already disrespectful, more so if it’s replicated a thousand times without your consent.
In hopes of understanding this further, we tried to create an AI artwork based on specific prompts:
Experiment #2
Prompt: “Picasso’s Guernica in the style of Van Gogh’s Starry Night with nuances of Gustav Klimt’s work”
This AI-generated image falls short of expectation as it is more of Van Gogh’s work and less characteristic of Picasso’s Guernica where the highlight should have been the bull. Even though the Van Gogh rendition is on point, it would have been more appropriate to have the elements within the Guernica painting as indicated in the prompt.
Experiment #3
Prompt: Partial Lyrics of Billy Joel’s “We Didn’t Start The Fire”
Rosenbergs, H-bomb, Sugar Ray, Panmunjom
Brando, “The King and I”, and “The Catcher in the Rye”
Eisenhower, Vaccine, England’s got a new queen
Marciano, Liberace, Santayana, goodbye
We didn’t start the fire
It was always burning, since the world’s been turning
To test the ability of the AI program in handling complex prompts and images, we took some lyrics from Billy Joel’s hit song, “We Didn’t Start the Fire” and used it as a prompt. The result is a little bit more conflicted and complex than imagined. With old style posters paired with distorted wordings and unrecognizable characters, the resulting image is quite puzzling. The AI image produced a vintage looking poster with unrecognizable words and inaccurate portraits. It would have looked better if all subjects were condensed into a single composite.
The Bottomline
In an era where everything revolves around time, speed, and convenience, artists are pressured to create their best work within the shortest possible time. People now gravitate towards apps and features that can help optimize their workflow in hopes of freeing their workload and saving extra time to spend for themselves. Filmmakers and game designers can greatly benefit from these kinds of AI tools as they can help them speed up the process of creating backgrounds, ambiance, and effects easily. For graphic designers illustrators and writers, AI can provide a good base image for concept studies and working models for projects.
Although the use of the internet and AI tools is easily accessible and available for some, it is not always the case. Some AI platforms require certain hardware specifications to help the prompts run smoothly while others are hard to launch in a browser alone. Luckily, with the help of platforms like Vagon, anybody or anyone from your team can maximize the benefits of AI platforms without burdening their existing equipment. Since Vagon doubles as a computer on the web, you can run whichever program you choose and run it on your browser or mobile as long as you are connected online.
Conclusion
AI or Artificial Intelligence seems to fuel the pipedream project to give people more time through machine automation; however, it seems that the machine still has a long way to go in understanding deeply the nuances and complexities of human thinking. Given that it needs a million iterations and a dozen permutations for an image to arrive at a certain output, using AI tools might be more time-consuming than traditional methods.
Compared to traditional ways of doing artwork, AI-generated artwork’s difference is that although the product might look good and easy to produce there is not much of a story behind it. So maybe, AI-generated images can work for a quick fix, like a calorie bar instead of a quick meal which of course, can hardly compare to a full home-cooked meals. Are we looking for a quick fix or a spark of inspiration? or are we more willing to take down the path of a slower but meaningful journey? The answer lies with our belief and the path that we wanted to take.
