AI in design has caused a stir amongst the design community, and in inspecting this new-found tool, a number of revelations have come to the surface. Prior to this we discussed how the AI-powered software, MidJourney, still in its infancy has gripped designers. Sure, there is room for improvement, but MidJourney has indefinitely broadened horizons and made the future of AI and design something that is keeping everyone on their toes. But where is this future heading, what’s the nature of this collaboration and mostly importantly, how far along is it?
This article is the final installment of a three-part series titled, ‘Creator vs Creation‘ that spotlights the budding relationship between AI platforms, like MidJourney, and interior and architecture design.
Creator vs Creation: Can AI platforms like MidJourney replace designers?
MidJourney as An Architect’s Drug
Marco Vanucci, the Founder and Principal Architect of OpenSystem Architecture described MidJourney as ‘Architect’s Drug’ in a social media post. In revealing why he said that, he says, “AI has taken the design community by storm. With robotics and digital fabrication, we could program machines to manufacture bespoke artifacts. Now, with machine learning and AI we are on the verge of a revolution in design culture. The images are intricate and compelling as they blend a range of reference images fetched from a database of over 5 billion images. With AI, we get, in the blink of an eye, what would have taken weeks, if not months, to generate. The images can then be selected to create variations so that designers can explore the ‘design space’ of the idea described in the input text. For this reason, it’s rather easy to fall into a ‘rabbit hole’ where one gets trapped in endless permutations as the software keeps producing variations that are consistent with the initial description. In this sense, the process is very addictive.”
He goes on to bring out that some people compare this process to sketching. He describes it as ‘a quick rendition of ideas that retain a certain level of ambiguity and approximation while holding strange and, at times uncanny, aesthetic qualities.’
At present, although there exist constraints, Marco is hopeful about the future. “For the sake of design work, these images cannot be built as they are only visual illustrations. There’s no original file to measure, extract shop drawings, or 3D print. However, it’s not hard to imagine how this technology can be improved or applied to more precise and clear outputs. Although still in its infancy, AI is set to change the way we think, design, and build architecture.”
Architects as Curators, and Not Artists
Marco brings offers a different perspective. According to him, designing and architecture have already been involved in extracting data and inspiration from already-present sources, so it might be possible for AI to become a vital aid in designing and even coalesce with construction. “Arguably, architecture has always dealt with the process of indexing, sorting, and selecting architectural data. The codification of the classical order allowed architects to ‘automate’ the design process, reproducing identical copies of an original idea so that architects could claim authorship over ideas while delegating the construction to others. With the digital turn in architecture, CAD/CAM technologies allow architects to design and make, at the same time. The file we use for designing can be transferred to a machine for making. Now, with AI we enter unchartered territories: the creative process is automated. Although there’s still no connection with making, how long before we witness the full automation of design and making? In this context, what’s an architect’s agency? Can we still claim any authorship over the artwork? Or, rather, do architects work as curators, selecting rather than drawing images? Some argue that no original art can be produced by processes of assembling pre-existing data/images. However, hadn’t architecture always been concerned with sifting, sorting, and blending architectural data? Is the question of style still relevant?”, Marco says in keeping his argument forward.
“Digital tools are reliable for producing precise repetition. However, AI-generated images are never the same. The same prompt will never produce the same results – similar to hand drawings and sketching, AI imagery possesses some of those analogic qualities that were precluded to digital works. Each image presents idiosyncratic traits, and a final stage of permutation can never be determined. This process is shifting the focus from designing a single artwork to types and typologies of artworks. Rather than drafting, architects might end up having to sharpen their language”, Marco concludes.
AI in collaboration with Computational Design – How real is the possibility?
Kevin Abanto, an architect who dabbles in digital artworks, NFT and Metaverse says, “It is possible for AI to replace manual 3D modeling workflows. Imagine that AI generates results in three-dimensions and you can edit it as you wish. However, if we talk about Grasshopper or Houdini, where they can include computational, generative, procedural design – it is still a bit far. Right now, AI-supported tools like MidJourney are raging over social media. They are not just being used in the production of images, but also in the post-production and scaling of 3D visualisations, in virtual reality, NFTs, metaverse, for research, and much more. By the look at this rapid growth, there might be a surprise in store – maybe the conventional workflows that complement each other so well today, might be completely replaced tomorrow.”
What to do about AI in the present?
“The AI are very powerful tools, they are changing everything we know about art and the creative process, and it wont stop at that,” says Hassan Ragab. “We should learn to harness this power to our benefit by learning how to control it, by being updated and aware of the technology. We need to be able to discuss a lot of concerns regarding the usage of these tools without prejudice or bias and with an open mind.”
On the other hand, Edward McIntosh appeals to designers to be a little more careful, ‘’AI text-to-image tools are a fantastic new development for our “image reliant” practices. Though, there is a danger that these tools will further separate built-environment practices like architecture from the actual physical environment and that future generations lose even more connection with aspects of our profession such as materiality, tectonics, structure and the interplay between them in a cohesive architectural response. But, for Metaverse world-building I think these tools are amazing. Probably we will see new architectural specialisms such as ‘metaverse architect’.”
Gauging from the experts perspectives, there seems to be a mixed consensus. While some remain strong supporters of the potential of AI, others seem to think it might not make too much of a difference to a designer’s current role. However, what all of them agree on is the brilliance of these tools and how it is definitely a feat in world of technology. Who knows maybe AI might just be the change the design community requires and cultivates a generation of architects and designers that can focus on optimizing design output over drawing quality.
Read more about MidJourney
Previous in series: Creator vs Creation: What the users say about AI-platforms like MidJourney in architecture and design?
Previous in series: Creator vs Creation: Can AI platforms like MidJourney replace designers?