AI Concept Art
Shepard's Tone Unplugged: The Artistic Rebellion of AI Concept Art
In this issue of The Brief, I delve into the world of AI concept art and its profound significance within the film industry. AI concept art bridges creativity and production, saving time and money. It swiftly generates visuals and reduces human resources. Yet, it sparks controversy as it challenges traditional artists' roles, raising questions about human creativity in this evolving landscape.
The Siren Song of AI
AN AXIOM TO REMEMBER: "The 10 Axiom," that posits when a technology enhances the efficiency of a given task by a factor of ten or more, it ushers in a sea change in corporate behavior, driven by the pursuit of profit.
In the context of AI-driven concept art, this axiom holds particularly true, as it elevates the efficiency quotient to an astonishing degree, often making the process 100 to 1,000 times more efficient than traditional methods.
Consider a film project with a $100,000 budget for concept art.
In the conventional scenario, this budget might yield 20 to 40 pieces of production worthy concept art, each requiring a turnaround time of 3 to 5 days, necessitating a skilled team of artists led by an art director. Sourcing such a team can be time-consuming, potentially spanning several months, and often results in stylized concept art.
Conversely, for just 1/10th of the budget, say $10,000, AI technology enables a single prompt artist to generate an astonishing 1,000 pieces of photorealistic concept art, with real actors in scenes, and with a turnaround time measured in minutes.
This advantage is so substantial that it's difficult to envision any future film production not leveraging the remarkable potential of AI-driven concept art, signifying a transformative shift in the industry's approach to visual storytelling.
Is AI Concept Art any Good?
THE DANCING ELEPHANT AXIOM -“It’s nice to see the elephant dance, but I wouldn’t take it to the ballet.”
It's a common occurrence for professionals faced with a new technology that has the potential to replace their roles to initially dismiss it as laughable. They often view it with a sense of disbelief, considering it a mere joke that such technology could ever supplant their expertise. However, what they may not realize is that this seemingly laughable technology can evolve and advance to the point where it indeed becomes a serious contender for their roles, reshaping industries and professions in the process.
In fact, the question of whether AI concept art is "good" or not is, in many ways, the wrong question to ask. The more pertinent inquiries revolve around the trajectory of AI concept art's development. We should be asking at what velocity this technology is advancing toward achieving artistic excellence, how long it will take to reach or even surpass human capabilities, and how much better it can become artistically, in terms of time and budget efficiency compared to human artists. The focus should be on the continuous evolution of AI-driven concept art and the potential it holds to redefine the creative landscape within the film industry, rather than seeking a binary answer to its current quality.
Midjourney: a case study.
A PATTERN TO WATCH: What intrigues me is the number of artists who initially experimented with Midjourney, encountered minor setbacks like generating images with six fingers, and hastily dismissed it as inadequate. Yet, it's remarkable to witness the substantial improvements in this technology over time.
Since its founding in 2022, Midjourney has been on an astonishing trajectory of improvement. Starting with Version 4.0, released in November 2022, the platform introduced an entirely new codebase and AI architecture, significantly enhancing its understanding of creatures, places, and objects (and perhaps most importantly - fingers). Subsequent versions followed at a remarkable pace: Version 5.0 arrived in March 2023, offering more photographic generations and matching prompts with precision. Version 5.1, released on May 4th, 2023, further improved coherency, natural language interpretation, and image quality. Most recently, in June 2023, Midjourney unveiled its Version 5.2 model, setting new standards in efficiency and image quality, while also showcasing its advanced responsiveness to parameters.
This is nearly a new sub version release every month
Compare that to other historical release cycles. The release cycle of Microsoft Windows, a cornerstone of the computing world, spans decades with major versions arriving every few years with only minor improvements, whereas AI release cycles are measured in months and contain massive improvements.
Today, many artists view AI concept art as useless in production because it cannot do many of the required elements of filmmaking.
Maintain character, wardrobe, and prop consistency
Control scene blocking and character body positions
Control camera position, lens, and camera movement
Lighting, environmental replication, and cinematography
But that’s changing rapidly. Given the remarkable speed of Midjourney's progress, it's not far-fetched to postulate that in the next few years and maybe even within months, the platform may address previously unsolved challenges within the realm of filmmaking. This includes scene blocking, cinematographer mimicking, shot list creation, character and prompt consistency, and even the transformation of 2D photorealistic images into immersive 3D scenes and movies.
A THEME TO REMEMBER: It's crucial not to dismiss AI after just one attempt, as the AI landscape is evolving at an unprecedented pace. In fact, the changes in the industry can occur on a daily basis. To illustrate, I dedicate 2-4 hours daily to researching, experimenting with, and evaluating AI advancements in the context of film. However, even with this focused effort, it's becoming apparent that staying current in this rapidly evolving field is a formidable challenge, and I find myself struggling to keep up.
Is AI Concept Art Stealing?
The question of whether AI concept art constitutes "stealing" raises a thought-provoking inquiry into the nature of creativity and learning itself. AI, such as Midjourney, derives its capabilities from extensive training data that includes the work of existing concept artists. It is important to note that humans have a long history of learning, copying, and building upon the work of others, a process often celebrated as gaining experience and improving one's craft.
A BIG QUESTION: In this context, the crux of the matter lies in whether on a key questions “is it ethically acceptable for a computer to mimic and iterate upon human-created content in the same way that humans do”.
This debate touches on the evolving relationship between humans and technology, raising fundamental questions about originality, creativity, and the boundaries of intellectual property in an increasingly digital and AI-driven world.
Shepard’s Tone Concept Art: How I’m going to test the boundaries.
First and foremost, I'll dive into the ethical, artistic, and legal aspects of integrating AI into our creative process. I'll begin with an open-minded approach, testing the waters without predefined rules, and then gradually establish guidelines to ensure artistic and ethical integrity while staying within legal boundaries. This exploration will encompass ethical considerations like the authenticity of AI-generated content, questions of authorship, and preservation of artistic integrity. Additionally, I'll navigate the legal terrain, carefully examining how AI impacts copyright and intellectual property. The goal is to gain a comprehensive understanding of the multifaceted ethical, artistic, and legal nuances associated with AI integration, all while aligning with our ethical principles and legal obligations.
Simultaneously, I'll conduct a rigorous assessment of AI's effectiveness in producing production-worthy concept art for Shepard's Tone. I'll subject AI-generated concept art to a series of tests, evaluating its capacity to meet the rigorous standards demanded by a professional film project. This evaluation will encompass aspects such as quality, creativity, practicality, and alignment with the project's overarching vision and narrative. The objective is to gauge whether AI can serve as a valuable and efficient tool in our creative process, significantly contributing to the development of concept art that meets the professional standards required for our film.
Furthermore, I'll embark on the exciting endeavor of streamlining our workflow by employing ChatGPT to construct prompts for Midjourney. This automation aims to enhance the efficiency of our concept art creation process, facilitating the seamless integration of AI technologies within our creative pipeline. I'll explore ChatGPT's potential in comprehending and interpreting the specific needs of each scene and character. Simultaneously, I'll address potential challenges related to maintaining creative control and consistency. This innovative approach aims to leverage the full scope of AI's capabilities while preserving the essential creative vision and authorship inherent to Shepard's Tone. Through these three critical phases, I aim to thoroughly evaluate, adapt, and refine our AI-assisted concept art development process, opening new horizons in cinematic storytelling.
My predictions as to where we will end up.
Amara's Law, often attributed to American futurist Roy Amara, states: "We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run." This law highlights a common human tendency to have exaggerated expectations of new technologies when they are first introduced, followed by a period of disillusionment when the initial hype doesn't materialize as expected, only to be followed by a more accurate understanding of the technology's long-term impact.
In this exploration of AI concept art, we've witnessed a dynamic interplay between human creativity and technological innovation. As we ponder the future of artistry in an era of rapidly advancing AI like Midjourney, Amara's Law serves as a poignant reminder. We often underestimate the profound impact of technology in the short term, while the long-term consequences are far-reaching. Similarly, the Dancing Elephants axiom cautions against dismissing seemingly nascent technologies. While it may be tempting to overlook the early stumbles and shortcomings of AI concept art, doing so could mean overlooking the remarkable potential it holds. The fusion of human ingenuity and AI's computational prowess has the power to redefine the boundaries of artistic expression and creative collaboration. Thus, artists and creators alike are urged not to ignore the dance of these evolving elephants, for within their rhythm lies a symphony of artistic possibilities that promise to shape the future of visual storytelling.
Tomorrow, I begin the sprint. I’ll post 3 sprint updates - Tuesday, Wednesday, and Thursday - and then post a retrospective at week’s end.
What is The Brief and Who should read it?
I release a weekly digest every Friday, tailored for professionals ranging from executives to writers, directors, cinematographers, editors, and anyone actively involved in the film and television domain. This briefing offers a comprehensive yet accessible perspective on the convergence of technology and its implications for the movie and TV industry. It serves as an efficient gateway to understanding the nexus between Hollywood and Silicon Valley.
Who am I?
I'm Steve Newcomb. Functionally, I’m a recovering Silicon Valley founder that is finally old enough to have a bit of care. I’m perhaps most recognized for founding Powerset— it was the largest AI and machine learning project in the world when I founded it. It was later acquired by Microsoft and transformed into something you might recognize today - Microsoft Bing. Beyond Bing, I had the privilege of being on the pioneering team that witnessed the inaugural email sent via a mobile device. My journey also led me to SRI (Stanford Research Institute), where we laid the groundwork for contemporary speech recognition technology. Additionally, I was a co-founder of the debut company to introduce a 3D physics engine in Javascript. I've held positions on the board of directors and contributed funding to massive open source initiatives like NodeJS and even the largest such project, jQuery. My experience extends to academia, having been a senior fellow at the University of California, Berkeley's engineering and business faculties. Recently, I ventured into Layer 2 internet protocols and assisted a company named Matter Labs in securing $440 million in funding to bolster their endeavors.
What am I doing besides writing these posts?
Typically, I allocate a year between groundbreaking ventures. My exploration for the upcoming project commenced in May 2023, and the sole certainty is its nexus with the film, television, SMURF, and AI domains. Sharing insights on my research endeavors helps me discern between feasible prospects and mere illusions. My hope is that for this venture, I appropriately consider the ethical and sociological repercussions.
If you are interested in contacting me, being interviewed, being helped, or yelling at me, my email is steve.e.newcomb@gmail.com.