Comparing DALL·E 2, Midjourney, and Wombo: UX implications



Original Source Here

Comparing DALL·E 2, Midjourney, and Wombo: UX implications

The intersection of experience design and AI for digital art: where are we headed?

By Aideal Hwa from Unsplash

The evolution of large language models (LLM) [2] is a tough act to follow. It is no longer all bark and no bite.

We have had “citizen data scientists” for applying analytics principles to every day data use cases, citizenship matter as a concept that ended up becoming ubiquitous and so widely accepted that new training emerged (to inform how to become one).

Are you a “citizen artist,” an artist, or other if you utilize artificial intelligence (AI) methods to create art?

I revealed in a previous post, which informed how we can commercialize digital outputs from DALL·E:

“Users get full usage rights to commercialize the images they create with DALL·E, including the right to reprint, sell, and merchandise. This includes images they generated during the research preview” [3].

Productization

Specific to Wombo and DALL·E, the opportunity currently exists to commercialize, from inception to going-to-market, such that we can create products, optimize them for the market, and sell them. Read some of those details here [3].

Before proceeding further, I will demonstrate in Part 1 how to activate outputs using these AI tools. Showing this approach is essential because there is absolutely never a good or correct method in how to activate these AI models. In Part 2, I will hone in on experience design and the intersection with digital art.

To quickly optimize the parameters to inform the models, I used [4] to curate the specific sentence to inform Midjourney. I revised Midjourney prompt (but tried to keep it as close as possible) for Wombo and DALL·E 2.

By FIsh God from

Part 1: a few cities rendered (and the how-to)

1. Capri, Italy.

The prompt for Midjourney (again, utilize [4] to understand the underlying infrastructure — it is essential to learn the language to speak on similar terms for each method):

Capri, Italy, during sunrise showing a sense of awe, Modern Art, by Claude Monet— v 3 — q 1 — s 2500 — stop 100 — ar 16:9

From the Author, Midjourney result for Capri, Italy

The prompt for DALL·E 2:

Capri, Italy, during sunrise showing a sense of awe, Modern Art, by Claude Monet, ar 16:9

From the Author, DALL·E 2 result for Capri, Italy

The prompt for Wombo:

Capri, Italy, during sunrise showing a sense of awe.

— The art style identified is Malevolent. No input image was provided.

From the Author, Wombo result for Capri, Italy

2. Palm Beach, Florida

Midjourney Prompt:

Palm Beach, Florida, during sunrise showing a sense of awe, Modern Art, by Salvador Dali — v 3 — q 1 — s 2500 — stop 100 — ar 16:9

From the Author, Midjourney result for Palm Beach, Florida

The prompt for DALL·E 2:

Palm Beach, Florida, during sunrise showing a sense of awe, Modern Art, by Salvador Dali

From the Author, Dall E 2 result for Palm Beach, Florida

I reran this prompt a couple more times; it appears I cannot evade the hand (other than entirely eliminating the majority of the prompt, a dependency likely caused by the use of “palm”.)

The prompt for Wombo:

Palm Beach, Florida, during sunrise showing a sense of awe

— The art style identified is S. Dali. No input image was provided.

From the Author, Wombo result for Palm Beach, Florida

3. Copenhagen, Denmark

Midjourney Prompt:

Copenhagen, Denmark, during sunrise showing a sense of awe— v 3 — q 1 — s 2500 — stop 100 — ar 16:9

From the Author, Midjourney result for Copenhagen, Denmark

The prompt for DALL·E 2:

Copenhagen, Denmark, during sunrise showing a sense of awe

From the Author, Dall E 2 result for Copenhagen, Denmark

The prompt for Wombo:

Copenhagen, Denmark, during sunrise showing a sense of awe

— The art style identified is Malevolent. No input image was provided.

From the Author, Wombo result for Copenhagen, Denmark

Part 2: experience design and implications

Digital art is no longer a dream as a wish the heart makes in secret. Through experience design, an environment can be created from scratch to activate the senses and emotions of the user to integrate them across the value chain [17]. This type of design as they relate to the rise in such digital artistry could potentially create a sense of immersion, which can help achieve a type of experience (how memorable, how intense?). By using experience design, we could create an environment to inform how users interact with their work (at a minimum to ensure that they walk away with a very specific experience).

The present, growing interest in experiential designs for digital artworks has been inevitable due to advances in AI. I have never felt that visual art is a passive experience where users observe work that is already complete. Still, with AI-generated artworks now permitted by capabilities like Midjourney and DALL·E 2, users can interact with the artwork directly in ways not possible in the past, permitting and allowing for a new type of engaging and immersive experience.

How can user experience designers [11] work with artists (or on those specific productization life cycles) on use cases, like incorporating elements of sound, visualization effects, or movement into products or solutions to influence the digital results?

A great deal of thought goes into creating an experience that will stay with users long after they part with the environment. Every element is carefully designed to achieve the desired effect (if an experience designer wants people to feel anxious, they might use dim lighting and suspenseful music to create an unsettling atmosphere. On the other hand, if they want people to feel excited or happy, they might use bright colors and amplify them with specific type of music.)

By Changqing Lu from Unsplash

In the case with DALL·E 2, it can be integrated to build that experience as a tool for creating prototypes and user flows. This approach in tooling can be particularly impactful in working out how users will interact with an eventual (deployable) product or service, and in identifying any potential bottlenecks or issues that need to be addressed.

A different approach for DALL·E 2 is to apply it as a platform for testing and iteration. By continuously tweaking and improving [13] the design of a product or service, experience designers could work towards ensuring that deployable products meet the needs of users. Ultimately, the goal is to create a desirable user experience [14] that meets (and exceeds) users’ expectations.

Generative design is a type of computational design where algorithms are applied to generate new designs [10] based on specified constraints and parameters. This approach allows for a vast number of potential designs to be created, outputs that can be filtered and refined according to the needs of the project. As an illustration, a designer may use generative design to create a series of abstract shapes to apply as graphics in an interactive installation. By tweaking the algorithms and parameters, different results can be rendered, giving the designer a versatile toolkit with which to experiment.

By tommao wang from Unsplash

Elements of experience design at the intersection of AI-generated art

If an artist wants to use DALL·E 2 to explore how users interact with art, should the elements of understanding human perception and cognition apply? In other words, the more we understand about users’ visual processing abilities and how they interpret different kinds of stimuli (iconography being one example), are we more prepared to design for interfaces that require people to make sense of digital content?

Research has shown that creating layouts that take advantage of cognitive load [6] influences users in how they maintain focus [7] on task at hand [5], summarized to inform how all elements in the interface we design are cohesively hinged [15] and demand (our, the designers’) attention. Additionally, effective feedback mechanisms may allow users time to adjust their internal processes to give maximal benefit from whatever information or objects they are viewing [8].

By incorporating these methods into workflows, like layout analysis techniques [16] based on cognitive load theory and consideration for appropriate feedback mechanisms, the opportunity may exist for experience designers to delve further into why a user would actually engage with digital artwork in the first place.

By Lance Chang from Unsplash

Parting Thoughts

Patterns are integrated or developed through industrial design [12] principles [9] to create visual interest or emphasis on certain areas of a prototype.

We have not even scratched the surface with 3D animation using present-day AI-assisted digitalization capabilities, especially to inform 3D printed design products.

Industrial designers could potentially apply DALL·E 2’s “edit image” function (working as of 8/3/2022) to create and MODIFY patterns. This capability is currently very rudimentary, but its present day form is, as I believe, a part of the data they may be collecting (from how we are using it to manipulate the rendered output) to inform its future, purposeful function. Specifically, its vector editing capability (using the DALL·E 2 tool) allow right now for highly precise adjustments.

Is there an opportunity for third-party (3P) integrations, software to software, to create seamless life cycles for design?

Please share your thoughts with me if you recommend any edits for this post or recommendations on further expanding this topic area.

Also, please kindly consider subscribing to my weekly newsletter.

References:

UX Collective community:

11. Lamenza, F. (2021, July 1). There’s something wrong with the UX Designer job title. UX Collective. https://uxdesign.cc/theres-something-wrong-with-the-user-experience-designer-title-6f6b1d06c262

12. Mühlstedt, J. (2021, April 5). The anatomy of a design trend. UX Collective. https://uxdesign.cc/trends-anatomy-d18099094df0

13. Fordham, M. J. (2021, February 7). A practical guide on how to choose and tweak typefaces for your design project. UX Collective. https://uxdesign.cc/a-practical-guide-on-how-to-choose-and-tweak-typefaces-for-your-design-project-7076f0505505

14. Silverstein, S. (2020, September 29). Usable, useful, desirable… and buildable? UX Collective. https://uxdesign.cc/designing-for-build-cc3869ac5d79

15. Fard, A. (2020, May 5). Why should product teams use wireframes more often? UX Collective. https://uxdesign.cc/why-should-product-teams-use-wireframes-more-often-60e34a2bc55

16. Moreno, L. (2020, December 3). Fundamentals of layout in user interface design (UI). UX Collective. https://uxdesign.cc/fundamentals-of-layout-in-interface-design-ui-3a9dba31f1

Mueller, S. (2020, July 23). Looping the consumer into the value chain. UX Collective. https://uxdesign.cc/looping-the-consumer-into-the-value-chain-62077eb5cf2d

AI/ML

Trending AI/ML Article Identified & Digested via Granola by Ramsey Elbasheer; a Machine-Driven RSS Bot

%d bloggers like this: