Day 1 – Create Your Own Stock Images – 30 Day AI Challenge 2024

Welcome to the 2024 Edition of the #30DayAIChallenge. This year, we’ll be focusing on making Generative AI more accessible for all. So follow along and give each challenge a try yourself, plus share your own thoughts and experiments with a community of like-minded learners and tinkerers!*

Week 1‘s theme is “Enter The Matrix, Creativity Reloaded“, focused on what you can do personally to tap into your creative spark, plus bring ideas to life.

Introduction to Today’s Challenge: Creating Stock Images

Why pay for Stock Images when you can create your own custom images that do the job better? Creating your own stock images sounds like a task reserved for professional photographers and designers, right? Wrong. With the rise of Generative AI, the power to create stunning, high-quality images is now at everyone’s fingertips. Today’s challenge is all about unleashing your inner artist and creating custom stock images without the need for a professional setup or years of design experience.

How Beginners Can Approach This Challenge

For those new to Generative AI, the idea of creating stock images might seem daunting. However, several AI tools have made this process incredibly user-friendly. Tools like DALL-E, Google Gemini, Bing Image Creator, Stable Diffusion, Leonardo.ai and others are great starting points.

Here’s how to dive in:

1. Choose Your Subject: Start with a clear idea of the image you want to create. Whether it’s a serene landscape, a bustling city scene, or abstract art, having a specific concept will guide your creation.
2. Pick an AI Tool: For beginners and those looking for free tools, Bing Image Creator, Adobe Firefly, Stable Diffusion, Google Gemini and Dall-E 3 (integrated with ChatGPT) offer an intuitive interface, while tools like Leonardo.ai provide more advanced options with detailed prompts. Most tools offer free credits or limited daily image generations to try things out.
3. Craft Your Prompt: This is crucial. The more detailed your prompt, the better your results. Include aspects like setting, mood, main objects, colors, and style.
4. Iterate: Don’t settle for your first result. Play around with different prompts and settings to see how slight changes can affect the outcome.
5. Use it Responsibly: Remember, the images you create should be used ethically, respecting copyright laws, personal rights and labelling / disclosure where necessary. Also check the usage rights associated with different platforms…for example some sites only allow for commercial use if you’re on a paid plan (in most cases still cheaper and better than buying stock photos).

For our test today, I decided to create a stock image for a destination marketing project, welcoming guests to experience the incredible culture and sights in India. Here’s what some of the tools above created with the following prompt:

“Photo showing an Indian woman in traditional clothing with folded hands and a welcoming smile”

Adobe Firefly:

Bing Image Creator (Copilot – Powered by OpenAI’s Dall E 3)

Stable Diffusion Web (there are various sites that offer their tech)

As you can see, most of these free to use tools do a pretty great job of creating realistic imagery, perfect to design your own stock photos and visuals to suit very custom asks.

Advanced Tips for Seasoned Users

For those already familiar with Generative AI, pushing the boundaries of what these tools can do is the real challenge. Here’s how to elevate your stock image creation:

1. Hybridize Tools: Combine the outputs from multiple AI tools. For instance, generate a base image with DALL-E and refine it with Photoshop AI filters or enhance it using Gigapixel AI or other upscalers.
2. Experiment with Complex Prompts: Utilize advanced techniques like style transfer, specific art styles (e.g., “in the style of Impressionism”), or even mixing elements (e.g., “a futuristic city with 18th-century architecture”).
3. Fine-Tuning: Use platforms that allow fine-tuning or training your own models on specific types of images to create truly unique stock photos.

Midjourney is still the “Rolls Royce” of AI Image Generators but there is a learning curve, plus you’ll need a paid subscription. You can get VERY creative and technical with the available parameters…or just keep things simple, based on your needs and proficiency. At present, you can access Midjourney image creation through Discord, though there is also an Alpha version of their web app in testing…this will likely be rolled out in the coming months more widely. Here are a couple of examples of the same prompt, explored with Midjourney…

Midjourney: Same, simple prompt without any additions

Midjourney: A more fine-tuned prompt to adjust Style, Aspect Ratio, etc.

Pro Tips

• If you’d like very prompt adherent image generation, I’d recommend using Dall-E 3 within the paid version of ChatGPT…given the Large Language Model (LLM) itself crafts the prompts, it does an excellent job of staying on point. The only drawback with Dall E 3 is that it doesn’t create very photo-realistic images of people, like Midjourney does.
• There are loads of tools that can allow you to create great, detailed prompts, including custom GPTs within ChatGPT.
• Seeking inspiration from others is a great way to level up your creations. For example, most websites feature user-generated image galleries, where you can also see what prompt was used…this way you can copy or tweak the prompt for style or content to see what you can come up with.

Wrap Up and Future Possibilities

Creating your own stock images (and visuals in general) with Generative AI opens up a world of creativity and personalization that was hard to imagine just a few years ago. As these tools continue to evolve, we can expect even more detailed and realistic images to be produced at the click of a button or by simply asking your favourite chatbot / tool verbally. The potential for customizing visuals for blogs, websites, presentations, and more is limitless.

If you give this a try, leave a comment and share your experience or creations! If you’re using other platforms, please tag your Facebook, Twitter / X or LinkedIn post with #30DayAIChallenge so others can find it too.

Till tomorrow…

Other Posts You May Find Interesting:

Gen AI Image Battle 2 – MidJourney, DALL E, Stable Diffusion, Leonardo and Adobe

The Lines Are Blurring – How Good Are You At Spotting Real vs AI Generated Images?

Generative AI – Watch Midjourney Quality Evolve Exponentially In Just A Year

Generative AI Magic: See How 15 Timeless Fictional Characters Look Like In Real Life

*Please note: Participation in the 30 Day AI Challenge is at your own discretion and responsibility. Always ensure that no sensitive personal information, confidential, or proprietary company data is shared. Adhere to all applicable local laws and company policies. Enjoy exploring AI responsibly!