Modern architecture prioritises profit over user preferences, neglecting the potential for beauty and well-being. Can AI and VR bridge this gap?

The potential of AI and VR in modern architecture

Modern architecture prioritises profit over user preferences, neglecting the potential for beauty and well-being. Can AI and VR bridge this gap?

Artificial intelligence (AI) and virtual reality (VR) applications are increasingly becoming a part of our lives and living environments. As more people spend more time interacting with them and, sometimes, immersed in the “metaverse”, the issue of the impact of the “digital” life on the real life of the user is becoming more relevant. In addition, data from user-system interactions can give us a glimpse into user preferences on a scale that was unattainable in the past.

Objective beauty

A substantial body of research now supports the notion of objective criteria for beauty, distinct from subjective stylistic preferences. These criteria are rooted in certain geometrical aspects commonly found in both pre-modern architecture worldwide and in natural environments. Our innate connection with the natural world predisposes us to find certain geometric features inherently appealing, aiding in our comprehension of scenes and often enhancing their perceived beauty.

Elements contributing to this “objective beauty” include fractal-like structures, nested symmetries, local contrasts, and other visually meaningful cues reflecting organized complexity. Studies utilising portable sensors or virtual reality simulations have investigated the physiological responses to exposure to different morphological features. Among their findings is an observed increase in stress response when individuals are exposed to geometries lacking these features.

Decades before neuroscience delved into these matters, architect Christopher Alexander observed a tendency among designers to prioritise tidy, hierarchical structures, neglecting the intricate web-like connections inherent in overlapping, reinforcing relationships termed “patterns.” Alexander proposed managing these patterns more effectively within a language-like framework of broader relationships, known as a “pattern language.” We now understand better how the experience of beauty stems from our nervous system’s response to such stimuli, with humans naturally drawn to natural geometries, and we can rank structures according to their adherence to such geometries, in a kind of “objective beauty scale”.

The “expert”/”non-expert” divide

However, many design professionals possess quite different and strong views on what constitutes beauty or ugliness in architecture. Architects assess architectural beauty by virtue of their adherence to “approved” versus “unacceptable” design styles, practicing a sort of “geometrical fundamentalism”. These evaluations often diverge from the biology-based beauty scale, and what architects currently deem desirable points towards lower values. With a few exceptions, urban infrastructure constructed worldwide following World War II diverges dramatically and fundamentally from what was built during millennia of prior human building activity. There is an evident conflict between architectural culture and common preferences.

How can AI and VR help satisfy our innate need for beauty? As engagement with AI and VR is increasing, it can be leveraged for benefitting user well-being. Additionally, data from user interactions can give us a glimpse into design preferences on a previously unattainable scale.

Alexandros A Lavdas

AI data hint

AI programs like DALL·E 2, Craiyon, DreamStudio, and Stable Diffusion can generate photorealistic images based on verbal descriptions, having been trained on vast datasets of online images and captions. For instance, prompting for a “beautiful building” yields traditional-looking (pre-modern) structures deemed high in objective beauty, while requesting an “ugly building” results in industrial-modernist buildings, low in objective beauty.

Figure 1. Images created by Stable Diffusion following the prompts “Beautiful Building” (top row) and “Ugly Building” (bottom row).
Credit. Author

Midjourney, another AI-based program specialising in artistic paintings, produces colorful, curved buildings reminiscent of Art Nouveau in response to the “beautiful building”  prompt. These findings are consistent with global opinions on beauty.

These AI programs learn from user feedback to generate images corresponding to popular perceptions of beauty and ugliness, reflecting popular opinion on a scale unimaginable for traditional surveys. By sidestepping aesthetic ideologies, it exposes a critical divergence from dominant architectural culture: text-to-image AI generators echo the long-missing voice of the people in architectural discourse.

The construction industry profits greatly from narrow industrial-modernist designs, leveraging significant investment in public relations to manipulate public opinion. This keeps the lid on the natural inclination for alternative environmental geometries.

Despite recent AI applications in design, these primarily address technical issues rather than user experience. This limited use of technology perpetuates existing design paradigms rather than challenging them.

It’s important to note that arguments grounded in scientific analysis do not advocate traditional design for historicism’s sake; rather, they center on a style-independent analysis of geometric properties. Analogous to grammar and syntax in language, this analysis doesn’t dictate what to create but highlights what design elements to avoid. Viewed through this lens, modern design often exhibits numerous “grammar and syntax errors.”

But what about virtual environments?

Digital aesthetics play a crucial role in various sectors, shaping new possibilities in art and architecture. Still, the metaverse has yet to captivate users due to its lack of biologically-based beauty. Addressing this could enhance user engagement and transform the metaverse into a therapeutic environment.

Asian-Pacific companies heavily invest in Metaverse platforms to escape physical urban dystopias, yet design styles often lack elements that could result in real engagement. Many recent VR designs reflect deliberate architectural choices influenced by current physical world trends and constrained by simplistic industrial forms. However, to further attract participants, virtual environments must prioritize objective beauty through biology-based design.

Consideration of what is called affordance in psychology, referring to what the environment offers in the broadest sense, has been neglected in favor of visual formalism; however, such consideration could enhance the user experience, and its implementation could be led by data collected from the general public.

There are positive signs from the film and video game industries; settings associated with morally positive elements tend to be traditional, while opposites lean towards techno-modernist styles. Architectural discourse often avoids discussions of beauty and ugliness, focusing instead on unrelated topics. Styles from the 20th and early 21st centuries dominate today’s construction, with abstract minimalism prevailing despite not aligning with the aesthetics of most people.

An intriguing scenario

This introduces an exciting potential utilising the metaverse for therapeutic purposes. In such a scenario, Metaverse headsets incorporating eye-tracking and facial expression sensors (which are already available) will constantly monitor signs of the users’ overall well-being. The feedback received can then aim to guide users towards a more positive emotional state.

Until now, altering the environment for health benefits has been restricted to simple elements like music and lighting, not real-time modifications to physical structures. However, virtual environments offer significant opportunities in this regard. In addition to the therapeutic effects, a plethora of research is expected to emerge, categorizing the visual appeal of places across a spectrum of engagement levels.

While a large-scale change in the physical built environment remains unlikely due to entrenched profit motives, the emerging Metaverse presents a different opportunity. Investors could realize that inhuman design fails to attract users, leading to a shift away from architectural “experts” toward user insights, as well as in the film and video game industries. A more appealing Metaverse can adopt old-fashioned, organic, and natural aesthetics reminiscent of traditional and vernacular cities or even explore new styles, always adhering to the geometries that our nervous system is tuned to perceive, with psychological and physiological benefits for the users.

Steps to promote psychophysical health through AI and VR

  1. VR designers should become aware of the findings from neuroscience and psychology on the influence of the visual environment on psychophysical well-being.
  2. VR designers should realize that the architecture of the virtual universe does not have to be held hostage to the limitations—the dictates of fashion, the imperatives of technocracy, and the economic and institutional power structures—that control and influence physical architecture.
  3. Architects, designers, and decision-makers in the physical world should not only become aware of these same findings. They should also listen to the voice of the people, as echoed through AI text-to-image applications.


Journal reference

Lavdas, A. A., Mehaffy, M. W., & Salingaros, N. A. (2023). AI, the beauty of places, and the metaverse: beyond “geometrical fundamentalism”. Architectural Intelligence2(1), 8.

Alexandros A. Lavdas, MSc, PhD, is a tenured Senior Researcher Neuroscientist at Eurac Research, Bolzano, Italy, an Assistant Professor and Head of Psychology at Webster University, Athens Campus, Greece, and a member of the Board of Directors of the Human Architecture and Planning Institute, Concord, MA, USA. He holds a PhD from University College London (UCL) and has worked at UCL and the Hellenic Pasteur Institute in Athens, and taught at the University of Indianapolis (Athens branch). In the past, he has worked extensively in nervous system development and regeneration, and in more recent years he is especially interested in examining elements of visual organised complexity, such as those found in nature and pre-modern architecture, and exploring their psychophysiological correlates.