(Left picture) - Person standing in large jar with lights in a starry night sky and on grass. (Right picture) - Yelling woman with black clothes and hair that looks like pain on the water in a cloudy sky.

Spotlight Series: Andrzej Zarzycki

At NJIT, we have many instructors that are implementing technology in innovative ways into their teaching. Of those, the Office of Digital Learning touched base with our first spotlight, Andrzej Zarzycki, Associate Professor in the Hillier College of Architecture & Design (HCAD) and asked him some questions on how he’s using Artificial Intelligence (AI) in the classroom.
Post Date: 
Thursday, April 27, 2023

All images in this article were created through Andrzej Zarzycki’s usage of AI.

(Left picture) - Modern Furniture design. (Right picture) - Finished wood that uses complex joinery for support.

Fig 1 Theme. Exploring architectural detailing (timber) and furniture designs allows for a quick prototyping of architectural assemblies and spatial components.

While Andrzej stated that current course offerings in AI within HCAD are limited, the instructors within the College are hard at work trying to introduce various aspects of AI design into the curriculum, particularly its generative capabilities. Andrzej shared that in the design studio courses, where deliveries can include areas from 2D/3D character design to interactive environments, generative AI can provide stimuli for design explorations.

There are already a number of AI-type applications that have been used in architecture for a long time, such as Grasshopper for Rhino, Autodesk Revit (Dynamo), or proprietary GANN models. Now there are newer tools, like Stable Diffusion, to facilitate generative designs through visual and textural inputs.

Andrzej notes that there is still some confusion and apprehension in the use of AI tools in the classroom. Particularly those recently developed for the general user market. For both students and creative professionals, the use of generative AI may feel like a “shortcut” due to the amount of quality original design. It parallels the discovery stage of design, where designers develop multiple designs to identify the most promising solution(s). Students can also use generative AI to experiment with various graphic styles for their design, giving them the opportunity to experiment. Andrzej encourages students to use AI tools to outsource some secondary tasks to generative AI to provide students with more time and focus on their primary design objective.

Fig 2 Theme. AI-tools allow for designs that combine speculative permutations of sub-components or unexpected characteristics. A steam powered typewriter or sculptural glass bottles emerge from the unusual combination of variables--a horizontal transfer of features akin to Edward de Bono's Lateral Thinking approach.  

A student may be interested in 2D character design and wants to tell the story using her 2D character. To tell a visual story, characters need to be contextualized within scenery/environment. Using AI to develop scenery allows the student to focus on her main personal interest – character design.

Andrzej Zarzycki

One direction proposed by Andrzej is the use of AI tools in the design curriculum to develop a consistency of outcomes – quality, style, and features across various scenes. To teach students how to use AI generative tools creatively and to not become a “victim” of developing new visuals without novelty and originality. While the general quality of creative/design work will improve through the use of AI tools, generic AI-quality will become a new baseline from which students and creative professionals will need to demonstrate more individualized and sophisticated applications to distinguish themselves. Andrzej believes that students and faculty need to know it to maintain a competitive edge.

Fig 3 Theme. Conceptualization and visualization of architectural forms with AI tools allow for tectonic variations, material substitutions, and style explorations.

This is where our (myself and my HCAD colleagues) curricular focus comes from. I don’t think we can compete anymore with AI on the basic productivity and generative abilities. We need to learn how to use these tools in areas they are not yet addressing effectively

Andrzej Zarzycki

Fig 4 Theme. Space-making by synthesizing various artistic styles like Anime and Art Nouveau and resolving them within three-dimensional volumes. This allows for both artistic and diagrammatic depiction of interior spaces.

When asked about students’ interest in using generative AI tools, Andrzej noted that courses that teach AI skills have a rather smooth adoption rate. In design studio courses, students choose their preferred software tools and there has been less interest in using generative AI. In fact, he has noticed more interest in ChatGPT over visual AI tools for assistance with Arduino or other coding tasks, since that is a less familiar skill for them.

When it comes to the how generative AI can change digital design in the future, Andrzej says it’s hard to overstate the impact it will have. The quality of products and productivity will increase. However, he doesn’t believe that this will reduce the demand for professionals dealing with digital creativity. Instead, it will change their skillset and the type of students that NJIT and the profession attracts. Instead of clip art or stack photos for presentations or business publications, AI will deliver customized and original content. People will do digital design as a sidekick, similar to how content production has moved from large corporate media to social media content creators, like Youtube. At this time, the faculty at HCAD introduce these tools more individually within their own respective curricula.

Andrzej believes that AI should be leveraged and the mindset should be switched to teach students how to use AI generative tools creatively, rather than becoming a ‘victim’ of new visuals without the originality or novelty.

Students and faculty need to know it [generative AI] to maintain our competitive edge.

Andrzej Zarzycki