AI Generative Art: Hype, Fear-mongering, and Practical Use

Due to the rapidly evolving nature of AI generative art and LLM’s - this entry will be updated accordingly

AI Art: The Doomsday Discourse and Practical Realities

Lately, there’s been an absolute tsunami of AI doomsayers, and quite frankly, much of it seems to stem from an anti-humanist mindset rather than from a place of critical thinking or rationality. Believe me, I would know—I’ve passively tortured myself by wading through countless opinions, subreddits, and clickbait articles that range from mildly irritating to outright cringeworthy.

The hysteria is often amplified by artists themselves—sensitive, introspective souls who communicate in ways that make them naturally wary of this new, flashy tool unleashed upon us by the Tech Broverlords. Their apprehension is understandable, even predictable.

But what’s truly puzzling is the peculiar glee some people express at the idea of artists losing their income, identity, or relevance. It’s as if they believe artists have just had it too good for too long. Right? Because, obviously, painting for a living is akin to rolling in riches, isn't it?

Is It Really Practical?

Despite the doomsday predictions from those outside the creative sphere, I’m yet to see how generative AI could be practically integrated into the daily grind of a professional artist. The creative process is surgical—it involves careful attention to detail and total control over those details. Control is precisely what AI lacks. It doesn’t engage in the creative process; it skips it entirely and spits out an instant result.

On paper, instant results sound handy. But creativity doesn’t work that way. It’s about iterating, refining, and responding to specific needs—whether for personal fulfillment or a client’s requirements. I’d gladly welcome AI assistance to offload some of the grunt work, but so far, I haven’t found any tool that genuinely complements my workflow. My role requires solving a variety of problems across different fields, and for that, I need flexibility and precision.

Take generative AI for architecture, for example. It’s often hyped as revolutionary, but anyone who’s remotely informed knows it’s laughably inadequate. Architecture is constrained by complex rules, codes, and regulations that vary by country and even by local councils. Feeding all that data into an AI to generate a viable design would be a Herculean effort, and the output would still be limited by site-specific challenges. Even for those churning out cookie-cutter designs, land conditions throw a wrench into things.

In-House Generative AI: A Smarter Approach?

One practical use case I’m actually in favor of is in-house machine learning. The idea is to train a model using proprietary material developed within a business. However, this becomes trickier the more complex the tasks.

Take Blizzard Entertainment, for instance. They’ve developed their own in-house generative AI, trained on their existing art. The AI can generate new concepts while retaining Blizzard’s signature style, making it useful for quickly producing rough concept sketches. Contrary to popular belief, real concept art is often quite rough and is produced rapidly in the early stages of design. Those elaborate pieces you see in art books and marketing materials? They’re the polished exceptions, not the rule.

In these early stages, teams focus on shapes and silhouettes. AI could be useful here by randomizing and reinterpreting these shapes, giving artists more to explore. But even then, the process still requires a human hand to curate and guide it. Unsurprisingly, artists are still better at prompting generative AI, ensuring that what comes out has some semblance of artistic intent.

Retaining the Human Touch

In Blizzard’s case, AI is trained on their own creative assets, allowing it to generate concepts that fit their unique style. This helps speed up the early stages of design by producing rough sketches and ideas quickly. However, AI remains a tool—human artists are still needed to refine those ideas, inject emotion, and ensure the final product fits the game’s narrative and vision. While the AI handles the grunt work, the real artistry comes from human creativity and judgment, which are essential for bringing any project to life with depth and meaning.

While the fearmongers keep shouting about AI replacing artists, the reality is a bit more grounded. Tools like Blizzard’s in-house AI might help shuffle through rough concepts, but they still require a human touch to sift through the noise and shape the final product. AI can spit out random silhouettes, but it’s the artist who breathes life into those lines, tweaking them with precision and purpose. So, while AI might be handy for knocking out a quick sketch, when it comes to the real heart of the process, the paintbrush still belongs in human hands.