”Art-ificial” Intelligence“
With the emergence of ChatGPT, as a text to text generative AI a spring of other Artificial Intelligence tools have been developed in its wake. Forerunners Stability AI‘s Stabile Diffusion Software, Midjourney (available on Discord) as well as ChatGPT’s very own DALL.E are Image Generation Tools which generates unique, new images from text prompts. Since 2022, diffusion technique programs have been the foundation on which Image Generation AI‘s are built.
Image Generation Tools
It is now possible to merge AI and Art without using code, the concept that you can create, design or produce visuals just by writing in natural language what you conceive in your mind has vastly become a marvel in the online content creating community. Image generation tools output surreal, abstract & imaginative images I am sure you have seen flooding social media for the past year. These images are a result of text descriptions (rather than instructions) which guide the ai to to generate images within a few minutes. Image Generation Programs have many creatives and art enthusiasts creating art using Artificial Intelligence to produce never before seen, one of a kind visuals. If you are like me with no-technical or design background and would like to create your own images with ai generated tools, read on to see recommendations on how to create images with AI generated tools; using programs’ that I have personally created with.
Leonardo AI
The first popular Image generation tool I used was Leonardo AI, which can be used to create production-quality visual images from text using prompts. This was all very new to considering that I am not an artistically minded at all, fortunately I had a friend who is a developer talk me through the what and how’s of Leonardo AI, text prompts and tokens. After I creating an account and signing in, you are greeted with a dark mode style platform with pink, purple and blue gradient buttons. A brief introductory guide should pop up (I recommend going through the steps). Navigate the dashboard to get to the Image Generation page, once there it should look like the image below.
Where there is a bar at the top that states, ‘Type a prompt…’ This is where your text description of what you want to create it written. The quality of the image generated depends on how clear and specific your input prompt is. Within minutes of entering the prompt, the AI will start generating images. This could take longer depending on the complexity of the tasks. Each image generation will cost you a few tokens, but Leonardo AI grants 250 tokens daily to use for the images. Once generated you can review the images and modify them with the model settings on the left of the page.
Every image generated gets saved into your generation history, you then have the option to download or export the file. Its important to note that Leonardo AI is not an AI but a software app which uses Stability AI’s Stable Diffusion.
Pollinations AI
Another web application which uses stable diffusion is Pollinations.ai. This website allows you to generate images from text instantly, but there is no way to control settings with your prompt. This is a basic form of using an image generation and its only use was because I didn’t always have paid version of chat GPT. Seeing as GPT 3.5 does not include image generation, the loophole around this was to use text prompts in the form of generative image URL by Pollinations AI to include images in the output on ChatGPT 3.5. The pollinations ai webpage itself looks like an internet art gallery of generated images but besides this feature its only use is for image url’s. If you want to know more about how you can generate images click here.
DALL.E
DALL.E is OpenAI’s built in text description to image feature. It is really good at generating images in GPT 4 but it does not generate letters & words very well.
To access DALL-E’s image generation features, a subscription to ChatGPT Plus or Team is required. It is not available with GPT-3.5; to use DALL-E in ChatGPT-4, you simple start a new chat, select ChatGPT-4, and enter a prompt starting with “Generate an image of…”
The fusion of human creativity and artificial intelligence opens up new horizons for artistic expression, challenging our traditional notions of what it means to be an artist and a viewer. All three tools have their strengths and are suited to different use cases.
If you’re just starting out or want a quick and easy way to generate images, Pollinations.ai is a great choice. If you’re willing to spend some time learning and want more control over the image generation process, Leonardo AI is a good option. And if you’re already using ChatGPT and want a seamless experience, DALL-E is the way to go. Happy creating!
Leave a Reply