Stable Diffusion has officially arrived; here are some approaches to be user-centric.
Dozens of organizations are cohering AI power for an intentional digital art competition. Who is going to win over the most users? How many people will use their services? Just how large a footprint may their implementations of large language models (LLM) for converting text to image make?
The aforementioned questions are ones that I currently hear many professionals discussing anxiously. From my perspective, it is not so easily obvious that user experience, such as the design of these organizations’ websites or software for their understandability or usability, maybe a ubiquitous topic of discussion or debate.
These text-to-image capabilities are inaccessible unless you make use of the provided software or visit the associated websites. Furthermore, it seems that the go-to-market speed is so rapid that businesses may be prioritizing going live with their capabilities first while concurrently building or optimizing the user experience.
Each of us has purchased a product that we later realized had insufficient testing and development before being released to the public. In contrast to those experiences, these AI solutions seem like (to me) they needed to be released, go live, and made accessible to users as quickly as possible.
I recently wrote about the go-live updates regarding Stable Diffusion: they want to operate differently; the industry is noticing how special it is. A Figma plugin is already underway to produce design solutions. Stable Diffusion is open-source: I can take the source code and “pretrain” my learnings to create very specific designs or outputs for my use cases.
Stable Diffusion has only been available to the public for about three days at this point. If you do not wish to program using code to generate solutions, DreamStudio is the hub of their product. To give you an idea, here it is.
This dashboarding solution is not just a text entry box like Dall·E 2, Midjourney, Wombo, and others.
Will Stable Diffusion pay attention to the fundamental principles of design?
“Users’ expectations about discoverability… are set by their prior experiences with all the other software they have used.”
How easy is it for users to find the features and content they need within a given interface?
Labels should be clear and concise, and they should use words that are familiar to the user. These are essential for the users to find the features they need in an interface quickly and easily. There are several more concepts to consider in the context of labeling. First, it helps users make a mental model of how the interface works. If labels are unclear or misleading, it will be hard for users to figure out how to use the interface well. Second, familiar language makes it easier for users to find what they need. If labels are too technical or jargon, users may have trouble finding the right content. Lastly, labels that are clear and concise to understand can help reduce fatigue and frustration. If users have to stop and try to figure out what words mean all the time, they will become frustrated with the system.
Also, menus should be well-organized and easy to use. Navigation should be easy to understand and make sense, using common practices when possible. Search options should be robust and allow for filtering as needed. Finally, error messages should be helpful and tell the user what to do next as much as possible.
“Generating feedback is typically less of a concern with physical devices because the user can feel when a button depresses or how far they’ve pushed a joystick.”
Imagine utilizing physical products to generate solutions without augmented or virtual reality; imagine the day when we can create solutions that appear as output instantaneously using augmented reality.
Visual Cues are essential: use colors, shapes, and animations to give users feedback about their interactions with the interface. For example, they could use a color-coding system to help users know what actions are available to them at any given time.
Audible Indicators are important such that they can use sounds to give users feedback about their interactions with the interface. For example, they might use a chime sound when a user completes an action successfully or uses a particular feature of the app for the first time.
Concerning haptic feedback, Stable Diffusion can use vibrations to give users physical sensations that match what users do in the interface design. If a user causes an error with the system, the tool can simulate error messages with this kind of haptic feedback.
“The notion of affordances implies that the user is matching his or her goal against the set of opportunities offered by the environment, that he or she directly sees what can be done to reach the goal.”
What is the relationship between the tool and the user, where the tool’s properties offer clues about its function? For example, a doorknob affordance might suggest that the door is meant to be opened by gripping and turning the knob.
Adjusting the values of each parameter is simplified by using sliders (as shown in the visual above). If the button labeled “Dream” is clickable, then the action of “submitting” is also a possibility. Additionally, I like how the brightness color is included in the slider since this allows the program to imply (by brightness) which parts of the screen may be manipulated by the user.
With the current state of the dashboard in mind, three considerations emerge. First, maintain consistency throughout the design by always using the same affordances for the same activities. Therefore, the UI will be more natural and straightforward. Also, keep things simple; don’t overwhelm users with too many flashy graphics or animations. Never lose sight of the fact that the primary objective is to convey to users exactly what features they possess. Finally, make it obvious for users to understand what each affordance is for by providing them with descriptive names and attractive images.
“The relationship between controls and their actions follows the principles of good mapping, enhanced as much as possible through spatial layout and temporal contiguity.”
To map anything is to establish a connection between entities. In UX, mappings are used to establish connections between components, making it simpler to convey the functionality of a product to users.
DreamStudio should be evaluated using many factors. Size and proportionality should be mapped to an element’s relevance or function. One possibility is to make the “Dream” button bigger than the input area next to it. Elements should be arranged according to their spatial relationships. Unrelated items can be farther apart than related ones. We think of hierarchy when organizing information from most to least significant vertically, but we may also apply it horizontally across content sections and nest objects within containers.
The following practices may help DreamStudio:
— Create user flows before low-fidelity wireframes to map out what screens and interactions are needed to complete a product task.
When creating future layouts, give likely conversion points more emphasis, such as larger size or higher placement. What’s with the bottom-of-the-page text prompt?
— Use color to provide significance beyond aesthetics; use tints and tones sparingly.
— Think about font pairings when choosing typefaces.
— Group like items together spatially or by consistent styling.
5. Conceptual model
“Conceptual models can be taught, but they are continually refined throughout a user’s interaction with a system, given its responses to the user’s input.”
A conceptual model is a representation of how users interact with a system. It can be used to design and engineer user interfaces, as well as to assess the effectiveness of an existing system.
Some examples of how Stable Diffusion might use a conceptual model in their work:
— To see how users use it (as a product).
— Predict users’ reactions to product modifications.
— Find usability issues.
— Create user scenarios that describe DreamStudio interactions, and develop wireframes.
— To construct wireframes.
When building a conceptual model, consider the system’s users. Personas are fictitious characters that represent user groups. Task analysis breaks down the steps of a system task. Also, consider how users now use the product and how they may react to changes.
User scenarios describe common interactions with systems from start to end. Wireframes are another crucial part of any successful conceptual model; they lay out each screen of an app or website so designers can see what to add.
To design effective conceptual models, undertake significant user research to understand the desired audience. Existing user users should be evaluated. Idea models should be dynamic, adapting to user feedback and changing behaviors.
UX includes every element of the user experience, from usability to appearance. A successful product can be distinguished from a failing one by its strong user experience; understanding users’ needs is the first step in creating a good user experience.
As for what comes after DreamStudio and Dall·E 2, it is 3D video animation. Everything that we are inputting into these models are being utilized to build features for the next wave of artificial intelligence in digital art.
When it comes to what comes after DreamStudio and Dall·E 2,, 3D video animation is the next step. Everything that we feed into these models is being put to use to develop features for the next wave of artificial intelligence in digital art.
If you have any recommendations for this post or suggestions for broadening the subject, I would appreciate hearing from you.
Also, here is my newsletter; I hope you will kindly consider subscribing.https://predictiveventures.substack.com
References are in-text embedded. Additionally:
1. OpenAI assisted with the development of this visual.