Zaha Hadid Architects, AI & the future of architecture | Patrik Schumacher

I'm a globalist. I'm a cosmopolitan. And I believe there's a unified world architecture.

Maybe sustainability should be stopped for a year. They should stop talking about social justice for a year or two or three. Let's stop fabricating at architecture schools with robots.

Well, yeah. I mean, definitely everybody is using it. You can't avoid it. And it's actually maybe co-creation is a good phrase, but it's maybe overselling it a little bit. But definitely for me it was not only always a productivity boosting but also creativity boosting. It helps to be more creative because you can explore more and you can explore it in ways which are not relying on what you can imagine but you're relying on processes which can dip into huge reservoirs of imagery and you can blend, compose, and combine references and associations through the prompting. So in terms of the image AI, there's a lot of creativity boosting. So we like that. At the same time we got the more rational process, many new tools. We have been hitting on this new tool called XFigura which is this kind of tool where you can design workflows, multi-author or multi-participant, and generate these panels and making links and tying in and having all the models of image generation more or less available to select from and string along into a workflow similar to Grasshopper. So that's been very potent. There's other new AI companies not only the big models but specifically architecturally oriented AI startup companies like Jendoai and XFigura. The other thing which I want to emphasize is the writing of programs and optimization tools in the form of web apps has been becoming easier and easier. So I was getting myself back into coding and using Cursor and generating web apps. But now with Gemini, it's been very efficient and good the way you can instruct programs and then let them run and then even have an AI analysis of the result and the way the program operates. That's just staging as well as any research you want to do, background research. So the kind of boosting of the rationality content of our work through these optimization tools but also just through learning quickly about materials and things, planning systems, etc. So there's an enormous boost of intelligence, rationality, cognitive capacity if you like, but also again creativity. So not only image AI is feeding in on all kind of levels. It helps us to push forward faster and in more sophisticated ways and be more compelling in visualizing as well. That's a big topic as well for us which is also cost effective and time effective. To animate our models and render them out with also with little video moves, it's something which really helps to understand what we're presenting. It helps even for ourselves to gain more conviction about what we're putting forward. So overall, yeah, it's been a fantastic journey and we're just at the beginning of it.

Oh yeah, absolutely the case. But it's also creation, I think it's also still which when we first talked it was inception and ideation was foregrounded and I would still foreground that too, of course. Then we have the kind of visualization and of course that transition from ideation into then we still have to model properly and develop the design of course and then to render that out again. And the clients are very much aware of what's going on and if you delay the modeling for too long and you keep showing them AI generated images which don't have an underpinning through a proper geometry, they also get suspicious. So they know that this is not totally trusted in terms of proportions, in terms of realization and so on. Everybody is clued in at this point I would say, including clients who appreciate that one can get out of the gate very fast but also look at more ideas in parallel, but then also getting wary that if one doesn't back it up through proper design in terms of 3D geometry which meets the criteria of the brief, site, etc., and you get proper visualizations. So that's why in the end for the visualization we have full 3D models. We still use rendering company specialists and I don't know how they work. They probably have integrated massively various AI tools too. And it's coming in on all levels.

Well, no, no, I don't think so. I think we're actually lacking a certain amount of convergence. That's what we need. We need to converge to a set of, bring everybody up to speed on the new set of fundamental values, principles, ways of working, methodologies. And AI can help with that. What we are suffering from now is a fragmentation of the discipline which shows up in city development, in urban development, that shows up as what I call garbage spill urbanization because every architect was just following his own kind of personal predilection style. There's many different styles to choose from and there's a kind of sense, there's no sense of cohering an urban fabric in a contemporary way because you could have neopostists next to neotraditionalists, neoclassicists next to minimalists next to some of us parametricism coming in, and that generates something which is indigestible, which lacks coherence, and nobody likes the result. The result is a sort of strange sameness out of too much difference. If you pile in uncoordinated, unrelated, pure difference, you get some kind of identityless soup which looks similar because we can't comprehend. It doesn't build up an identity. An identity needs to build up an order where elements relate to each other and then not just a heap of diverse rubble, let's say. And everybody feels there's no identity anymore. It's just fragmentation, incoherence, which signals that we don't have a rigorous discourse. A discourse needs to weed out bad recipes, backward recipes, self-indulgent ways of working. And we need to critique each other and critically engage and debate the pros and cons of various sets of principles and paradigms of ways of working. For instance, what I believe is that when we come to developing an architecture on an urban scale, we need to allow the buildings to be resonating with each other, inter-affiliating and multiply affiliating. We need to converge, and the lack of convergence is what holds us back. Of course, converging to the most sophisticated, most tooled-up, most rationally defensible, and most versatile set of repertoire and methodologies. You need to converge like in the 1960s, finally the whole world, wherever you were, whether it was in India, in Africa, across Europe, the US, Canada, anywhere around the world, everybody had absorbed and finally the lessons and competences of modernism. It took 40 years to do that, but it was achieved around 1960.

I would have that same criticism. I agree with that, but it doesn't mean that you can therefore remain within your little vernacular structures, masonry with small windows, etc. Yes, it was relatively more similar to each other's solution, although they also had that element when they went into the tropics or into Indian zones. The brise became deeper. There was an attempt to affiliate a little bit in the Middle East. The solutions looked a little bit different, but it was much more standardized and the solutions were more repetitive, not only across the world but within the city with each individual building. And yes, this is no longer our time because modernism existed in a mechanical age. There were no computational tools available to society and this was a mechanical reproduction society. Now we are living in a computational age where everything is much more fluid, dynamic, and you can also in terms of production, reprogram the robots, rescript the service software service offerings. But again, we need to converge and everybody needs to be at the height of the civilization and contribute on the level that the civilization requires, and not be locked in backward modes and moods of working. And I think the diversity will be much higher than modernism achieved. Because we have the spirit, we have much more tools, analytic but also empirical tools of gathering data, of feeding data into our solutions and let them drive our solutions, also local data, local materiality, and also social, analyzing and understanding social institutions and behavioral patterns more intricately. And it's more diverse, societies are more diverse. They share a lot because we all live in that knowledge economy, post-Fordist network society if you like, but there are also a world division of labor and different cultures. The cultures are not necessarily Indian versus Chinese versus European, but the cultural difference will be the creative arts versus the finance industry versus legal profession, architecture versus engineering. They're quite different cultures, we work quite differently. These are the cultures which in some regions and parts of city are specializing in different contributions to the world economy, whether it's more mining and extractive and agricultural based societies, they also have their science and bioscience, or whether it's tourism based or educational based industries versus creative arts versus finance. So these are the different cultures which make different districts of cities or if some of these industries dominate in particular cities, it will shape these cities. And of course there are still some kind of vernacular cultural residues and they might also be productive in being functionalized for particular aspects in the world division of labor, but they would fade. What also stays of course is the climatic conditions which are regionally very different and the material resources, substrates. These things will stay and will also imprint themselves strongly. Tectonism is not just like modernism saying everything is now reinforced concrete. We're looking at stone, we're looking at timber, we're looking at compression only, 3D printed masonry, tensile structure. So it's more diverse, more varied than modernism. Modernism is the age of mechanical mass reproduction. So I'm a globalist. I'm a cosmopolitan. And I believe there's a unified world architecture. And the regional adaptations are delivered and built into the DNA of that global paradigm. That unified, universal, global architectural paradigm of parametricism, tectonism, has inbuilt that regional adaptive differentiation already. But to celebrate an architecture which remains fragmented where some people are still working in ways which one could have worked 100 years ago, that's very limited and inhibiting. That's not allowing these communities and these parts of the profession to prosper. And as soon as major projects come up in such a country, they will be snapped up by those who have been at the height of the world civilizational architectural competency. We will snap up all the cherries. So that's not helpful. And if the local architectures only do the kind of those who are still in only semi-developed states, this will be a dying business. From all over the planet you could have great architects. They might specialize in some particular aspects of their own region, but everybody should have a chance to rise to the top and build everywhere. Of course, the problem is it's easier for Europeans to develop in China than the vice versa, because there is nothing to do in Europe and everything is to be done in China. So everybody will move in there and they're very welcoming. And similarly I start to see that happening of course in India, but I also see it in the Middle East, which is also a wonderful playing field for global architecture. But at the same time I have to emphasize we need to be much more critical with each other, as well as self-critical, open ourselves up to criticism. And that is peer criticism first of all, most importantly in the field, in the discourse, in the discipline. And then of course we also expose ourselves to the criticism and selection criteria of our clients, clearly. And we have market pushback, market proliferation. But then I believe in discourses and markets to intercept, and the discourses allow us to anticipate, to critique ourselves. These discourses will bring up the leaders and then the markets will not fully understand why are these figures leaders, but they will trust our discipline that we self-select our leaders and therefore the best gets to succeed in the marketplace. That will be a healthy architectural culture. It's a global culture unless we're talking about backward regions who have so many problems that they can't develop a proper functioning prosperity engine. Then they have other issues. It's not something architecture could solve. As far as I'm concerned, we as a field we're in charge of the urban and architectural and design discourse and competency which we need to develop and bring literally everywhere, because that's just empowering and that doesn't lead to sameness. The problem is the incompetency leads to sameness, or it leads to trivial differences which are only, yes it's different because it simply hasn't caught up to what could be done, and therefore it's wasting resources.

So far, no. Look, this could have been said about CAD and BIM and all these wonderful tools of the computational revolution which made us massively more productive. I mean I definitely would say factor 10x in terms of the amount of things we can do in less time, which delivers much more sophisticated designs and much more clarity in communication with clients because they get early on a photorealistic rendering and videos of what it would be like. They can appraise whether this was the right kind of environment for their company, for their customers, but also knowing the geometry and costing of all of this and feasibility. So the planning intensity and sophistication and therefore quality of decision-making has massively ramped up through this productivity gain, but it hasn't reduced the number of architects and professionals. It has not. And the value of the design relative to construction has gradually increased, not massively but is increasing despite these enormous productivity gains. So I would think similar things will happen to architecture because architecture, yes of course if you were only talking about ideas competition or only about getting something viral onto Instagram, yes, there you could have bots or AI systems taking over to a large extent. But architecture is much more deep than that, there is much more complex and serious work. And for that the author, and also collective authorship through a firm and the network of consultants, is what counts. And also leadership within that. Whether the team and all the contribution is a co-creation between various tools, I accept that we as designers we're more and more developing systems and selecting systems and workflows rather than making individual design decisions. We're making systems setup decisions. But that means we're still authors on a higher plane. And in the end, authorship is basically taking responsibility, standing up and saying this is what we should be doing and here's why. And even if these arguments have been fed through an AI or tested with an AI discursive element, the creative ideation came in, many things coming in, structural calculations have been anticipated by topology optimization, another form of computational intelligence, and machine learned decking out of all the furnishings and space configurations. All of that doesn't take away that we are standing there as a team of authors with a figurehead and leader presenting and taking responsibility that these are sound systems, sound processes. And that is authorship. And of course there's no way that this could ever be substituted for. It could be substituted for when you want to have your own kitchen relayed out in a relatively simple way, yes, there's self-creation, self-co-creation through end users, consumers. But then again, one of our colleagues will design these systems. We're working participatory design systems where we generate modules and the combinatorics and we invite end users to plug and play. So this design is co-created, but we are the system designers. And that's the new role, a role of higher potentiality. But I believe we're still having a world population growing, but if not necessarily the total numbers of people growing massively, we have massive urbanization and upgrading our standards of living. So in the next coming decades there's an enormous amount of potential for architectural development. We need these tools to deliver to all of that and then we're not even starting to talk about the metaverse. Because once you put that into the equation, the amount of design which is required to bring that home is going to be a massive uplift in productivity which makes all our lives richer and more prosperous and more effective because we can communicate and navigate the world of online interactions much more real time, much more with veracity and vividness. If that comes through, the amount of design work required is gigantic. It's an avalanche of work coming to the design disciplines that we actually would need to gear up with all these AI tools and procedural ways of working to deliver that. But I don't think any of us are going to be substituted and cast aside and shifted to taxi driving. By the way, taxi driving is going, if you go to California. But all the better, there's other more fertile things to do. I'm not the one who believes that all these technologies deliver unemployment. They just deliver a more humane and more interesting and challenging and self-directed kind of employment. Who wants to clean up AutoCAD drawings?