AI’s Impact on Animation

With a career spanning three vital realms of animated projects — creative, production, and technical, ranging from shorts to episodics, interactives to features — I bring a discerning perspective to the ever-evolving world of animation. This diverse experience allows me to offer a balanced viewpoint, separating genuine transformation from fleeting novelties and unfounded fears.

On the technical front, I’ve had the privilege of designing, architecting, and contributing to numerous production pipelines for various film and TV studios, both renowned and emerging. Observing the industry’s rapid shifts in tools and methodologies has endowed me with invaluable insights. Even today, I remain a sought-after resource, offering guidance on how to structure animation production processes that are not only equipped for the present but also flexible enough to embrace the future.

Lately, the question I’m most frequently asked pertains to the impact of Artificial Intelligence (AI) on animation and the broader creative arts. AI has elicited a combination of wonder and trepidation among artists, with some heralding it as “Animation 3.0” and others fearing it as a threat to an art form with roots older than film itself. The truth, as with any change, likely resides in the middle ground. The real question that looms is where this transformative journey will ultimately take us.

This article marks the inception of a series — a journey into the realm of AI’s impact on creativity and animation. It is still the beginning, and I look forward to delving deeper into this fascinating intersection of technology and artistry with you.

Let’s start with some basics…

A Quick AI Primer

Artificial Intelligence (AI) can be defined as the “simulation of human intelligence in machines.” While this may sound daunting and fuel concerns of AI overpowering humanity, at its core, AI relies on algorithms (code) that process input to produce an output, similar to any computer program. What sets AI apart from conventional code is its capacity to learn from input, and make decisions or predictions, mirroring human intelligence’s dynamic problem-solving abilities.

Within the broader field of AI, Machine Learning (ML) is a specialized subset. ML focuses on constructing algorithms that can “learn” from data, identify patterns, and make informed decisions or predictions. Initially, these algorithms produce random or “noisy” results, but as they receive more data, they refine their patterns and decision-making processes, improving accuracy over time. In essence, ML models enhance their performance with relevant input, much like how artists refine their craft by studying the works of masterful predecessors.

Note: For a deeper dive into various AI techniques, I’ve penned a companion article outlining three common methods: GANs, VAEs, and Diffusion Models. Follow the link to explore these tools’ utility (and limitations).

Addressing the Fear

Now, let’s confront the most common fear: that AI will take over, rendering creative professionals obsolete (assuming it doesn’t destroy the planet in the process). This fear is entirely understandable, as individuals dedicate their lives to mastering their craft, only to contemplate being replaced by machines capable of instantaneous work. It’s evident in the remarkable creations from tools like Dall-E, SDXL, or Midjourney, which generate astonishingly “creative” pieces of art, often in novel and unexpected ways.

However, we must consider how these images are generated. Users direct AI by telling it what they want. Known as a “prompt,” it can be a text-based description or a set of images indicating the desired outcome. If the initial output is promising, users can instruct the AI to continue in that direction, generating related images.

The question then arises: where does creativity reside in this process? If AI can potentially replace an art department, what’s next for more complex fields like animation?

The Reality

Indeed, AI is a powerful tool, yet its an immature tech who’s true potential unfolds under the guidance of human creativity. Any potential “takeover” of tasks by AI will be a gradual process, as we are still in the early stages of the technology’s development. This extended timeline offers ample opportunity for adaptation to the new landscape and exploring with experimentation to see where it can go.

While AI yields remarkably convincing results, it remains far from flawless. Trained eyes can spot imperfections in generated imagery that may elude the average viewer. These imperfections, while acceptable at certain stages of production, fall short of the standards required for final output. In essence, AI functions as a valuable tool within an art department, akin to rapidly generating rough concepts during discovery. Its effectiveness, however, hinges on clear direction from an art director.

Although AI has been in use for some time, its recent surge in accessibility has democratized its application, granting access to individuals with minimal entry barriers. Now, anyone with a computer and an internet connection can submit their prompt to a web or chat interface and receive generated output. While this technological democratization fosters a wealth of innovative ideas, it simultaneously inundates the landscape with subpar concepts, posing challenges for talented individuals striving to distinguish themselves amidst the influx.

The most fitting analogy for AI in such creative applications is that of a “book-smart” junior artist. These AI models have absorbed vast amounts of information from the web and can proficiently draft requested content. However, they lack the experience and contextual understanding to fully grasp the underlying artistic choices and nuances, underscoring the continued importance of human guidance in the creative process.

AI, akin to an artist refining their craft, assimilates knowledge from every new input and adds it to its feedback loop. User selections essentially serve as evaluative scores, much like grades or critiques in an educational context, allowing the AI to align its responses with the user’s prompt and approved output, thus enhancing its overall performance. The prevalence of free access to many AI platforms stems from the desire to amass substantial data for continuous improvement. Similar to burgeoning artists, these AI models may not always perfectly align with an art director’s vision initially (and some go way off with “hallucinations”), but through iterative guidance and refinement, they can progressively enhance their capabilities and adapt to specific creative requirements.

The Impact on Animation

AI undeniably looms on the horizon for animation, but its impact varies across different roles. As of this writing, AI’s effect on animation remains relatively limited. Most AI tools primarily target text and still images. The 2D “animation” produced by AI mainly consists of morphs — smooth, interpretative blends between frames, not performance interpolation. Although this can be visually effective in certain scenarios, it represents a highly restricted form of animation. However, it provides a glimpse of where the industry is headed.

So, who is most affected? To my friends and colleagues at Aardman, styles like stop-frame animation remains relatively untouched due to its limited reliance on digital technology (thought that can vary). Conversely, traditional hand-drawn (2D) and Computer Generated Graphics (3D) animation will continue to experience the impact of AI, each in its unique way.

Let’s explore how each is influenced:

3D CG Production

The realm of 3D CG production encompasses various industries, from feature films to games, architecture, and simulations. The impact of AI varies depending on the style and context. For instance, there are already tools capable of generating lifelike motion using AI. Cascadeur, for example, excels at creating realistic animations with physics-based behaviors.

The individuals most immediately affected may well be those working in performance capture (aka Motion Capture, a field some argue falls outside traditional animation). Tools like Cascadeur can streamline the creation of generic and customizable motions. Some argue that this can be a positive development, allowing actors to focus on the expressive aspects of their performances rather than repetitive tasks like walk cycles. However, it then shifts the expectations for actors in this space, potentially impacting the talent pool. It’s important to note that, historically, the introduction of motion capture into games and film led to an increase in animation jobs, as more projects adopted the technology. So even when leveraging AI solutions, human input (an animator) is crucial to guide the process.

In the world of animation created by an artist on a computer (like with Disney, Dreamworks, Sony Animation, etc.), AI’s impact takes a different trajectory where AI has long been integrated into the pipeline. FX departments employ AI to generate advanced and realistic simulations. Layout tools have emerged to expedite scene dressing through predetermined asset palettes (commonly referred to as “smart scatters”) and textual descriptions. In these roles, artists remain in control of the final output, making necessary adjustments to ensure alignment with their creative vision.

For animators, AI offers a consistent impact across both 3D and 2D animation. AI will accelerate the interpolation of production-grade animations, enabling animators to transition from key poses to final output more efficiently, while maintaining the character’s arcs and weights. If they are dissatisfied with the result, they can introduce new keyframes and interpolate the animation. This shift will allow animators to collaborate more closely with directors on key performance elements, reducing the time spent on laborious in-betweening.

Traditional 2D Animation

Historically, hand-drawn 2D animation has been labor-intensive, as each frame is meticulously crafted one at a time. AI’s influence will be most pronounced in the cleanup and in-betweening phases, where an animator’s keyframes turn into fluid animation. AI systems are already demonstrating their ability to transform rough drawings into on-model final frames. As animators submit their initial keyframes, AI can swiftly generate near or final quality, in-between animations. With continuous exposure to an animator’s work, AI can learn to create more precise in-between frames that match the character’s style and performance.

One notable concern is the potential impact on entry-level positions, typically assigned to junior animators who typically hone their skills while refining the work of senior animators. This role has traditionally served as a fertile training ground for emerging talent. The narrowing of this path, combined with an unabated influx of creative individuals, will likely lead to the emergence of alternative career trajectories. Artists may need to demonstrate not only their animation skills but also their proficiency with AI tools to ascend the ranks.

In markets where outsourcing to specialized animation studios is common, AI holds several advantages over these studios, primarily in terms of speed. AI can rapidly produce output and operate without the limitations of time zones or language barriers. For example, key animators in France could submit their work to an AI and receive results almost instantly, instead of waiting for assistant animators to complete tasks, which may take hours or days. Realistically, AI will prompt outsource studios to rethink their business models, either by integrating AI themselves or transitioning to original content creation rather than work-for-hire services.

Change is the Only Constant

Change is indeed the only constant in the world of animation, and AI’s impact is unmistakable, though our journey is far from its conclusion. Every innovation marks the beginning of a cycle that involves evolution, adoption, and adaptation. As artists, our role is to remain vigilant, closely tracking the evolution of technology, embracing it when it aligns with our creative needs, and skillfully adapting it to our craft.

In the realm of technology, there’s a concept known as Jevons’s Paradox, originally proposed by William Stanley Jevons in 1865, which holds true in today’s tech-driven world. It underscores how as technology becomes more efficient and accessible, our utilization tends to increase. While some may view this as negating the benefits of efficiency, others see opportunities arising from the newfound ease of access to resources.

When individuals inquire about AI’s impact on jobs in animation and the creative arts, my response is rooted in the inevitability of change. AI represents a significant transformation on the horizon, with a preponderance of positive outcomes. Animation may require fewer individuals for specific processes, potentially rendering it a more cost-effective medium capable of delivering scalable quality. Alternatively, it may open novel creative vistas for those who engage with it, taking the art form into uncharted territories.

Throughout my career, I’ve been witness to rapid transformations in technology and media, from the shift from physical models to CG in renowned films like Titanic, Fifth Element, and Dante’s Peak, to the transition from hand-painted cels to digital ink and paint. In recent years, I’ve observed the emergence and impact of OpenSource, OpenUSD, Omniverse, and the mainstream acceptance of Virtual Production, thanks to powerful platforms like Unreal and Unity. In each case, these transformations were heralded by discernible signs, affording individuals ample time to adapt and find their place in the evolving landscape. AI follows this same pattern; it’s on its way, so maintaining a vigilant and open-minded stance is imperative. The innovations I’ve witnessed have consistently expanded opportunities rather than limiting them. In all instances, human creativity has remained the driving force propelling us toward exciting new horizons. Thus, my paramount advice is to continue learning and exploring, for that is the path to both mastery and innovation.

Reach Out & Connect

If you are presently immersed in the intersection of creative applications of AI within a professional context, or if you have a suggestion for a topic I should cover in this area, or simply some feedback on what I am writing, I would like to hear from you. Follow and message me on X/Twitter (at renderbox) or LinkedIn. Whether you or your team are actively engaged in developing innovative techniques or applications, creating new animated series, or producing feature films, your insights would be invaluable contributions to my ongoing series of articles, where we aim to provide an in-depth and easily accessible exploration into the continually evolving realm of Creative AI.