Midjourney-ing through the memories of a city and its architecture with Hassan Ragab

It has been nearly a year since David Holz's Midjourney shook the global creative community, amidst numerous debates across industries. Over the past year, the AI platform has surpassed the known potential of artificial intelligence in the creative domain and though it ventured into a world where programs such as Dall E and Stable Diffusion existed, its user-friendly interface brought forth a newer, easier, and ‘addictive’ mode of translating texts into visuals. With a user group that varies from curious AI fans to professionals from creative industries, the platform went from creating interesting imaginary visualisations to being a tool for designers and architects navigating the design process.

However, this also opened a potential inlet for designers of the non-western world, challenging the knowledge gaps of AI. As more and more visuals created in Midjourney surfaced, people started getting curious regarding artificial intelligence’s understanding of non-conventional styles of art and architecture. Following up on this attention, designers around the world jumped on the wagon and started exploring imaginative scenarios, merging traditional architecture and modern styles. However, for many of them, the curiosity was short-lived to a few renders, for some a series, and for others, it laid the foundation for identifying a whole new style of architecture.

Hardcore fans of Midjourney architecture might recollect visuals of modern adaptations of Egyptian architecture from the early days of Midjourney, which garnered mixed reactions. Most of those who weren’t familiar with Midjourney, then, wondered about its architectural practicality and necessity. Some argued about the imperfections of traditional architecture documented and some were impressed by the intricate detailing. This led to more designers coming up with experimentations of traditional architecture of different regions—Indian temples reimagined in a different era, Islamic architecture explored in terms of fluidity, biophilic architecture blurring the lines of built and nature, and many more.

Among the many visuals was a rendering which reimagined the trams of Egypt as a city. Though it seemed completely unrealistic, the visuals captured the intangible culture of the city through a conceptual image. The series titled, The City is a Tram, was a collection of images by Egyptian designer Hassan Ragab, with a description that read, "Trying to portray a version of 1900s Alexandria in my own (and of course my dear AI's) way. The Tram is an invaluable part of Alexandria's modern heritage, an iconic landmark penetrating ever slowly through the very heart of the old city. Providing a shelter to its very people, a daily commute dancing slowly while taking you to your destination.”

We traced back to the beginnings of Ragab's creative practice on Midjourney and he had a long list of styles, regions, concepts, and thoughts, that he had explored over the year, varying from underwater furniture design to high-density fluidic cities. Among the initial and consistent users of Midjourney, Ragab has experimented with a multitude of possibilities on the AI platform, Egyptian architecture and Egyptian cities being his favourite, owing to his roots in the country.

An architect by education, Ragab thinks that Midjourney helped him transcend the limitations of his profession’s practicality. In an interview with STIR, he mentions, “I have studied architecture but I didn't practice conventional architecture. I worked in several fields under the umbrella of architecture or design such as exhibition design, construction and furniture design. I am really interested in the visual aspect of the design and I have actually been challenged all over my years of experience to apply new ideas. It's always been very limited because my job required a specific result and I can’t expand much of my imagination around it."

"When I found Midjourney—it was the first platform that I actually used—it has been really addictive. The main reason for this was that I had no limitations on what I can try to create and the short time and ease with which imaginations can be turned into visuals. It has really helped me understand what I really like and understand what influences me. For example, it helped me connect better with Alexandria, a city in Egypt, where I was born. I don't think I would have generated an artistic piece about Alexandria if it wasn’t for Midjourney. I practice in the field of computational design and always tend to go towards the more advanced technological stuff and not the historical one. Midjourney builds a bridge between the advances of technology and history for me. And I find myself in between them. I think it's really empowering how I am trying to search or seek visual identities through that connection," he tells STIR over a video interview.

In his project, The City is a Tram, Ragab recreated his memories of the city and explored the possibility of a city that grows from within. The artist narrates that his intention to pursue the series lies in exploring AI's knowledge of traditional cultures of non-western countries and his confidence in knowing more about the city than the datasets.

However, while pursuing this, he was faced with a major constraint, one that is widely discussed when it comes to AI programs that use text-to-image commands—language. “The first thing I put in was Alexandria and I was really frustrated because it didn't understand the place. Midjourney thought it was a woman, an Arabic woman wearing a veil. So it was a challenge to translate the city into AI’s language. So, I kept working on that prompt for about three or four weeks, trying to add layers to it such as a building in Alexandria, the Alexandria Library or an Art Nouveau building in the city," Ragab shares.”

While Ragab shared the anecdote on Alexandria, we were curious about his thoughts on the debate surrounding language being a constraint. “English is my second language, so sometimes I struggle to get the right words. I see other people's prompts and am fascinated by their rich vocabulary. I always think if I could have this good vocabulary, it would have really influenced my work in a better way. But I think it's very similar to when you design anything. If you are designing or painting, without AI, you have to have a certain visual library in your head, which is like a different visual language. If you haven't seen enough paintings by Van Gogh or Paul Gauguin, you wouldn't be able to understand impressionism and you wouldn't be able to produce impressionism because that in itself is a language.”

I think language is always a barrier in the technology that we are using but it's not only the languages that we speak but also the languages that we see. – Hassan Ragab

Beyond his inquiry into Egyptian architecture and Islamic geometry, Ragab extends unravelling a multiverse of surrealistic interior spaces. Contrary to the earthy tones and shaded palettes of his architectural and urban works, his interior designs adorn colourful worlds of different materials, textures, and dynamism. An interior space inspired by crochet, another like a colourful cave, and another that reminds one of a playful candy house, are a few of his experiments. However, Ragab doesn’t confine himself to carefree whimsical styles but tries to merge architectural styles, famous artists, and architects in his works too. With designs that appear to be influenced by the natural intricacy of Spanish architect Antonio Gaudi, the fluidic transitions of Zaha Hadid or the direct representations of Art Nouveau, Ragab’s AI art blends characteristics of different inspirations to birth a unique style of design.

As he moves from Midjourney to finding the possibility of merging different AI platforms and how they can help in designing better for the real world, Ragab shares that Midjourney widened his thought sphere. “Midjourney opened a partition in my head where now everything is possible, a million infinite number of ideas that are out there for me to explore. That's how I see the usage of AI. The greatest potential of these tools is that they provide us with an unprecedented way to look over new vocabulary, new images, or new combinations, and they are happening so quickly in a manner that no human being will be able to keep up with it.”

In his series, The Cairo Sketches, Ragab dives into Islamic Cairo architectural facades, through 3D fractals. In an interesting introspection, many patterns seen in the series appear to be real-world jaali or second skin designs, widely used in parametric design. The series that unveils Islamic design symmetries extends to the viewers an evident connection between traditional practices and computerised design. Though history seconds this, in the fast-paced world where technologically advanced humans argue that machines are better than manual efforts, Ragab’s explorations serve as a reminder. A reminder of how much of our future lies in the repetition of history.

While debates on how AI platforms overrule artists of the world and the limitations of its creative output continue to occur even after a year, the program seems to be expanding its possibilities. Though different sides of public opinion still prevail at a strong ply of tug and war, one cannot ignore the newer discourses these AI results are bringing forward. As we investigate the harms of AI and scrutinise the popularity of these platforms, should we also be actively questioning the results they bring?

Ragab talks about his reason for creating one of his works, “The main exploration for these cities like Cairo or Alexandria is a way to examine the potential of AI images in historical context. It’s something that I can relate to because I was born and raised in Egypt, so I know these cities. I know the buildings really well and I can really relate to that and I am trying to look for a visual experience that I can relate to.”

There are things people know more about than AI, in such a context should we re-examine why certain things don’t fit the infinite mind of artificial intelligence? Should we consider what parts of our knowledge AI is truly conquering—the knowledge we develop or the knowledge we import from other sources? If so, should we first talk about the source than the intelligence it creates?