When generative AI first emerged, a familiar dread crept in. As a writer, the thought of algorithms replicating creative work felt… unsettling. Artists, I imagined, felt a similar tremor of fear as AI began to conjure increasingly realistic and stylized images. The question wasn’t *if* things would change, but *how drastically*.
But here I am, still employed, still crafting sentences and shaping narratives. And surprisingly, my trusty Photoshop remains an essential tool, used daily even amidst this technological shift. The initial panic has subsided, replaced by a pragmatic understanding: generative AI, while impressive, is far from a replacement for human skill.
The truth is, generative AI isn’t perfect. It excels in very specific scenarios, often requiring significant manual refinement. Despite my modest Photoshop abilities, I consistently find myself surpassing AI’s output – at least for the tasks that truly matter. It’s a powerful tool, but one with distinct limitations.
AI truly shines when tackling tasks outside my skillset. There are things I simply *can’t* do in Photoshop, and that’s where AI steps in. It’s also ideal for rapid prototyping and brainstorming, where absolute perfection isn’t the goal. Quick concepts, initial ideas – these are areas where AI’s speed is a genuine asset.
I’ve used it to create evocative concept art for roleplaying games, designing digital props and even crafting faux sci-fi interfaces for my players. Photoshop’s Generative Fill feature, for example, is a remarkably efficient healing brush and clone tool, quickly patching gaps or subtly expanding an image. These are small victories, but they demonstrate AI’s potential.
However, my enthusiasm stops short of grand artistic endeavors. While AI can generate sprawling landscapes, I have little need for such creations. It’s technically impressive, certainly, but ultimately irrelevant to my workflow. The core issue lies in AI’s fundamental inability to truly *understand* meaning or intent.
Even when I attempt to leverage AI, its lack of accuracy, precision, and adherence to specific instructions proves frustrating. It often feels like a sophisticated tech demo, lacking the reliability needed for anything beyond fleeting novelty. The results are often… close, but never quite right.
I recently attempted to create a character portrait for an Alien RPG campaign. The character, a corporate operative, needed a subtle visual cue: fingers splayed in “W” and “Y” shapes, representing loyalty to Weyland-Yutani. A seemingly simple request, yet AI utterly failed to deliver. Repeated prompts, countless iterations – the results were consistently wrong.
Hands faced the wrong direction, fingers were miscounted, the desired shapes were never achieved. After twenty minutes of fruitless effort, I abandoned AI and recreated the hand myself in Photoshop. Cloning a finger, adjusting the lighting, blending the layers – the entire process took just five minutes. It was faster, more accurate, and ultimately, more satisfying.
Perhaps AI provided a starting point, a rough sketch to build upon. But the final product was entirely my own. The AI’s output was no more useful than a random image plucked from the internet. More often than not, manual editing proves to be the quicker, more efficient path.
The central problem with current generative AI is the time investment required to correct its flaws. It often takes longer to fix AI-generated content than to create it from scratch. The infamous Coca-Cola Christmas ad serves as a stark example – a costly, flawed creation that ultimately required human intervention to salvage.
For those unfamiliar with tools like Photoshop, AI can offer a helpful starting point. But we’re not yet at a stage where manual artistry is obsolete, even at the hands of an amateur. The human touch still matters, and often, it’s the deciding factor.
There are countless small tasks I perform daily in Photoshop – resizing images, adjusting lighting, tweaking contrast – that AI simply isn’t suited for. It’s inefficient to rely on AI for these basic functions, even if it’s technically capable. Why expend valuable resources on something I can accomplish faster and more accurately myself?
These aren’t just matters of convenience; they’re also environmental concerns. The energy consumption and water usage associated with generative AI are significant. Being mindful of these costs is crucial. I’m not anti-AI, but I believe in using technology responsibly and efficiently.
I suspect that within the next decade or two, AI will evolve to the point where it can replicate my skills with sufficient accuracy. Perhaps my niche will disappear entirely, swallowed by the tide of automation. But I don’t believe we’re there yet. For now, AI remains a tool, a supplement, not a replacement.
I use AI for fun, for experimentation, for generating inspiration. It’s perfect for placeholder art in game design, or for quickly visualizing ideas. But when it comes to critical work, anything demanding precision, or anything I can already accomplish myself with ease, I’ll always choose the human approach. I’ll trust my own skills over an unreliable algorithm prone to “hallucinations.”
