Day 1: AI and creating images using Adobe Firefly
Welcome to Day 1! Today we are focusing on using image and text prompts to create images.
Today’s post is authored by Ben Turnbull, Senior Lecturer in Performance Technology and Design at London College of Fashion, UAL.
Adobe Firefly is an artificial intelligence system for generating images from scratch that have been created by Adobe. This means that it will be integrated into Adobe’s tools and will coexist with Photoshop, Illustrator and InDesign. It has been trained to generate images based on the description of what you want to see.
Adobe Firefly has a few differences that set it apart from other AI tools such as DALL-E , Midjourney and Stable Diffusion. Many AI models were trained on images scraped from the internet* with no regard for copyright. (More on AI and copyright to be shared in a few days!)
Where Firefly becomes a more powerful and consistent tool is when you combine your text prompts with reference images that communicate more of what it is you wish to see – or create.
You can use additional tool options within Adobe Firefly such ‘generative expand’, which builds out an image. Or ‘generative fill’, to select a portion of an image and then replace that content. For example, you can have an image of a person and then provide them with a hat.
* You can check if your own work trained an AI system by using Have I Been Trained?
How do I access it?
Go to Adobe Firefly and sign in with your UAL (or other institutional or personal) credentials.
If you haven’t used Firefly before, you may want to consult Adobe’s guidance for help and support.
Alternative tools
If you don’t have easy access to Firefly, you may want to try Stable Diffusion. There is a basic free plan available.
Activity 1
Create new iterations of images and themes based upon text and image-based information.
- Take a photo using your phone of a nearby object. This could be a mug, pen case or a piece of art.
- For this activity you want to choose ‘Text to Image’ in Firefly.
- Enter a text prompt into the prompt bar. I am going to go with “a large umbrella with a pattern of smaller rainbow cartoon umbrellas on the umbrella top.”
- Find the settings bar which includes a section called ‘Composition’ and ‘Style’. In both sections you have the option to upload your own image.
- Try uploading the photo you have taken to one or both sections. I uploaded a photo of a metal drinking cup.
- Composition will take the visual shapes in your photo and attempt to apply it to your prompt.
- Style will take the visual aesthetic of your image and attempt to apply it to your prompt.
- You can also use the reference images included in Firefly for this, but, when using it with students, the point is that they can work with documentation and references from their creative practice.
- Do not change the prompt. Just choose “generate” to get the new results.
- Now you can experiment with different images and changes to your text prompts. Try altering the other settings and visual styles you can apply to your prompts.
The first result showed where the outline of the cup is behind the umbrellas. So, I tweaked the filters.
I added a style, and this made the umbrellas shrink to fit on top of the cup in a pattern.
Discussion
Join us in the Teams space to share your responses and creations. If you don’t have access to the space, email us at teachingexchange@arts.ac.uk for the attention of Hannah.
- How many prompts did it take to reach your desired outcome? What kinds of prompts did you use?
- How do you think tools like this can be applied in your teaching? In your students’ learning?
- Do you think tools like this will become more industry standard?
- Was there any unexpected bias, stereotypes or benefits in these images?