A woman runs down an alleyway, a huge explosion behind her.......... As you can see, that is a pretty boring image for my first attempt at a prompt, and absolutely not what I had in my mind! She ain't running, and that's no huge 'explosion' behind her! However, as you can see, with plenty of patience, I did get there eventually. It's not perfect (AI still has a LOT to learn when it comes to hands and fingers!) but my desired outcome was achieved.
Read on to see how I learned to make better AI images.
When I first stumbled upon AI image generation, I was intrigued. The idea of creating art through algorithms and machine intelligence was fascinating, but as I eagerly pushed my first button, reality struck. My initial outputs were… let’s just say they were less than inspiring. They were plain, sometimes bizarre, and honestly, a bit disappointing. But as the saying goes, “failure is the stepping stone to success,” and my journey into the realm of AI art was just beginning.
The Initial Struggle: Underwhelming Results
Many people's initial attempts are usually characterised by generic prompts like “a beautiful landscape” or “a futuristic city.” More often than not, the results are images that lack character and vibrancy. For me, I quickly realised that this was a learning curve, much like any other creative endeavour. AI image generation required not just creativity, but a deep understanding of how prompts influenced outcomes.
I had entered a world where the precision of language mattered significantly. The AI's interpretation of my words was crucial, and my simple prompts were leading to mediocrity. I needed to dig deeper, experiment more, and embrace the subtleties of language.
Discovering the Power of Prompts
As I immersed myself in the world of AI art, I started exploring the idea of crafting better prompts. I read articles and watched videos, and what I learned from them is that the quality of an image is intricately tied to the specificity of the request. The more detailed and imaginative the prompt, the more engaging the results.
For example, instead of saying “a forest,” you can begin to experiment with multi-dimensional prompts like “a mystical forest bathed in the golden light of sunrise, filled with ancient trees and mystical creatures.” The transformation in the images will be astounding. Once I started to expand on my prompts, my art began to reflect depth, mood, and emotion.
Templates: A Game-Changer
While experimenting with prompts brought a noticeable improvement, I quickly discovered another valuable tool: templates. Templates (sometimes called LoRAs) can be thought of as add-ons to your prompt. They supercharge the AI with new capabilities and specialisations. You can for instance add templates to get more details, add characters, or create certain styles. By using templates, I could fill in certain fields and gather inspiration for my artistic direction without starting from scratch every time.
For instance, I found several templates that encapsulated various aspects:
- Setting: Where is the scene taking place? (e.g., underwater, in an alien landscape)
- Time of Day: What time is it? (e.g., dawn, dusk)
- Mood: What's the overall emotion? (e.g., serene, ominous)
- Elements: What key features should be highlighted? (e.g., animals, structures)
- Artistic Style: What's the vibe? (e.g., surrealism, impressionism)
Using these templates, along with some tweaking, my prompts became rich tapestries of ideas, and my image outputs began to reflect the imagination I poured into them. It was like wielding a magic wand; my simple requests could now conjure vivid worlds.
Embracing the Aesthetic: Continuously Learning & Adapting
As I continued my exploration, I developed a keen eye for aesthetic details. I learned to consider colour palettes, textures, and even the emotional weight of compositions. Engaging with other creators and sharing experiences became instrumental in this creative journey. I discovered trends and techniques that were being employed by more seasoned users and adopted those into my practices.
One important takeaway from this process was understanding the iterative nature of creativity, especially in AI. What worked once might not work again, and that’s perfectly normal. Each generation can be a unique adventure, and that unpredictability becomes part of the excitement.
Conclusion: The Art of AI is an Ongoing Journey
Today, I can confidently say I’ve transformed from someone who generated plain, uninspired images to someone who can generate fantastic works of art that resonate with emotion and allure. It was a journey paved with experimentation, learning, and discovery, reflecting the very essence of art-making itself.
In a world where technology meets creativity, the possibilities are endless. I look forward to continuing to challenge myself, further refine my prompts, explore new styles, and perhaps even inspire others to take their first crack at AI image generation. If there’s one takeaway from this evolution, it’s that there’s always room for growth—both for the artist and the AI.
So grab that prompt bar, embrace your imagination, and step into this exhilarating landscape. Your fantastic creations are just a few better words away.
Add comment
Comments