UNIT 2 PORTFOLIO
UNIT 2
PROPOSED REALITIES OF AI IN ARCHITECTURE
Artificial Intelligence (AI) can, in the right application, provide a more dynamic, informationally dense analysis of data for different iterative tasks. This process, implemented over a substantial period of time, can produce vast data sets in which different factions can analyze, develop, and eventually define software for commercial and individual use. These programs, as an example, can be used to mimic tasks in the commercial market that otherwise might be seen as repetitive or inefficient to meet more economical quotas. In many ways implementing these types of AI interface into select iterative practices seems like a time-saving and economically efficient way forward; however, there are certainly negative implications for this method as well. Design practices, like architecture, already have to contend with the infringing presence of AI, and with an ever-growing interest in the possible applications it could have for the field. Many questions are being raised as to the possible future realities AI could pose. Out of all these inquiries, one stands out as being particularly dire, a true homage to the ending of a design era, one that as a practice has lasted for hundreds of years and will surely be cut short by the insertion of AI design tools. While this is a melodramatic, foreordained point of view held by a select group, this statement still bears some truth in this figurative death sentence, invoking the question, Can there be a future reality in which artificial intelligence is implemented within the discipline of architecture to develop a symbiosis of the technological and the human aspect, or will AI surpass the need for designers, deeming architects an artifact of a now irrelevant past?
It is first necessary to understand what implications these types of programs can have on a practice like architecture. The process of design in many ways stays relatively the same throughout all disciplines: there is the initial problem, and what follows are the proposed solutions, often falling into a single iterative artifact, all pertaining to a proposed solution to that first problem. This is typically the job of the designer, someone who uses their skills and broad knowledge of the field to efficiently and effectively solve the problem; however, this methodology can be collected over time and eventually mapped out, essentially giving the designer a possible solution without even the slightest physical design or personal iteration. This hints at a reality where the human aspect of the design process becomes a less substantial factor in the development of the project. In Duo Dickinson’s article “Will Architecture in the Future be a Luxury Service?”, a point similar to this is made with reference to the exponential speed in which services are required in today's markets: “People still love growing food and making buildings. Faster, cheaper, easier solutions, via technology, may minimize the human hand in making things in all aspects of our culture.” Using comparisons from other disciplines, like food production, Dickinson highlights the broad nature of this spreading phenomenon and the possible outcomes that could affect more hands-on, repetitive practices. The article further illustrates the shift in design-related technologies’ evolution toward a possible architecture-less discipline: “In architecture, new technologies will inevitably mean that building consumers and contractors will have more ability to bypass architects entirely. Of course, some clients have always prescribed a design and the architects responded to their directives. But now, the next generation of software will respond directly, with even less need for an architect’s input. Data, once coordinated by architects, will be made seamlessly accessible and useful without the cost, pretense and liability of humans, especially those with a design-based education.” This is a possible reality for architects, one that can already be witnessed with small-scale, iterative design processes within the profession. This is further amplified when you take into account the level at which the majority of architects design. Small-scale residential architecture would be the first of the design levels to be affected by this shift in computing advancement, with AI being able to contemplate a multiplicity of simple design iterations in a matter of days--any human intervention would seem irrelevant in terms of both economics and efficiency.
This is only a single reality for the immersion of AI into the discipline of architectural design. There are in fact multiple paths forward in this evolution of design, many of which have a stronger constitution between the human and technological components. There are a variety of different opinions on the burgeoning topic of deep-learning computational machines, all of which hint at, if not shout out, the seemingly inevitable future of AI in replacing the human aspect in the digital and physical world. While there are negatives to this proposed reality, there are in fact working models of human machine symbiosis already happening today. In the article, "I Don't Really See AI as A Threat: Imdat As on Artificial Intelligence in Architecture,” the author, Michael J. Crosbie, discusses the many positive interactions that designers can have with artificially intelligent workflow machines: “With new AI-driven software, a designer might provide a host of constraints, say, a chair made out of a particular material that can hold 300 pounds. The software could generate hundreds of optimal chair variations that the designer could choose from and develop further.” This acts as a perfect example of the AI processing a design problem and, instead of taking the project away from the designer, expediting the initial iterative process, thus saving valuable time and money for both parties. This type of application can be used throughout the field, and being coupled through pre-existing design tools, like Autodesk Suite or Rhino modeling software, could greatly improve output performance of all design projects. This process can be derived simply from deep-learning machines’ ability to catalog data, most of which has already been stored on cloud-based platforms now made available to software companies that are building information databases for their AI projects. This allows the system to compare millions of possible examples and outcomes for a specific design question and problem using artificial levels of computational intuition, a topic that is also noted in Crosbie’s piece: “ Deep-learning machines could decipher patterns in architectural design that architects have intuitively or intentionally created over the years.” Crosbie also states, “AI could potentially expand such patterns to include not only functional and programmatic concerns, but also socio-economic, ideological, geographic, climatic, or other patterns that shape the built environment. With such a resource, AI-software could assemble the best patterns for a given problem into new compositions.” These new, relatively boundless parameters give architects the tools necessary to tackle even bigger problems than before, all while saving the necessary resources to do so.
In this reality, architects have limitless access to centuries of design practice, with a newfound precision in the development and perfection of dynamic, culturally aware social architecture in tandem with an expedited workflow, thus saving time and resources, all thanks to a symbiotic relationship with new AI-based design tools. Even with this template being possible in the near future, AI will inevitably evolve past this point as programs for computational learning surpass the need for any and all human intervention. This ideology spawns from the basic need to advance and improve exciting technologies. Even in disciplines like architecture, the physical ways in which structures are produced have been changing since the very first shelters were being erected thousands of years ago. This, in conjunction with a multiplicity of other evolving relative contexts, such as population growth, climate control, and famine, means that architecture needs to keep moving forward in the built environment as well as in the computational environment to meet projected future demands. In the TED talk “The incredible inventions of intuitive AI,” the presenter, Maurice Conti, explains the very real prospect that, with a very small level of human interaction, AI systems can design and now build more complex objects than ever was thought possible, using an example of a bridge being constructed over a canal in Amsterdam, by artist Joris Laarman and his fabrication firm MX3D. The goal was “to generativity design and robotically print the world's first autonomously manufactured bridge.” Once the team finishes designing the bridge with an AI system, robotic 3D printers will construct the bridge in its entirety--no human interaction required. This type of design fabrication process will be increasingly necessary with growing demands for faster, stronger, cheaper architecture, and, with AI development not showing any signs of slowing in the considerable future, we are sure to see a steady evolution in these self-learning/building systems.
All of these proposed realities place the architect in a world consumed by AI-entangled workflow and mechanical construction methodologies, all of which limit the architect’s ability to function as a designer or possibly even exist within a discipline. As AI grows within the discipline, the true potential of this computational tool will eventually be realized. Since AI is, relatively speaking, in its infancy, this level of pessimism is naturally expected among concerning parties, while, in truth, the development of AI software is still very much within our control. The TED talk, “Can we build AI without losing control over it?”, presented by Sam Harris, speaks to this level of understanding, stating that if software like the ones proposed here are developed in a sensitive and precise manner, a symbiotic relationship can, in fact, develop between AI and humans, in multiple disciplines, such as architecture and other design/construction practices. Articulation of our intentions will be critical for the evolution of artificial systems. In our usual hurried nature to evolve faster, we forget that we must also remember to evolve smarter, not cutting too many corners and inevitably weakening human involvement throughout the built environment. This would act as the final proposed reality of AI integrated architectural design--one in which we still need to work toward a more computationally humanistic system, where both parties can design, evolve, and prosper together, with AI acting as our extended consciousness in the solving of our world’s very real, current conditions and problems.
SOURCES
Dickinson, Duo. "Will Architecture in the Future Be a Luxury Service?" ArchDaily. August 02, 2018. Accessed September 17, 2018. https://www.archdaily.com/899483/will-architecture-in-the-future-be-a-luxury-service.
Crosbie, Michael J. ""I Don't Really See AI as A Threat": Imdat As on Artificial Intelligence in Architecture." ArchDaily. September 30, 2018. Accessed October 24, 2018. https://www.archdaily.com/902978/i-dont-really-see-ai-as-a-threat-imdat-as-on-artificial-intelligence-in-architecture.
Harris, Sam. "Can We Build AI without Losing Control over It?" TED: Ideas worth Spreading. June 2016. Accessed October 24, 2018. https://www.ted.com/talks/sam_harris_can_we_build_ai_without_losing_control_over_it#t-855839.
Conti, Maurice. "The Incredible Inventions of Intuitive AI." TED: Ideas worth Spreading. April 2016. Accessed October 24, 2018. https://www.ted.com/talks/maurice_conti_the_incredible_inventions_of_intuitive_ai?language=en.
PROPOSED REALITIES OF AI IN ARCHITECTURE
Artificial Intelligence (AI) can, in the right application, provide a more dynamic, informationally dense analysis of data for different iterative tasks. This process, implemented over a substantial period of time, can produce vast data sets in which different factions can analyze, develop, and eventually define software for commercial and individual use. These programs, as an example, can be used to mimic tasks in the commercial market that otherwise might be seen as repetitive or inefficient to meet more economical quotas. In many ways implementing these types of AI interface into select iterative practices seems like a time-saving and economically efficient way forward; however, there are certainly negative implications for this method as well. Design practices, like architecture, already have to contend with the infringing presence of AI, and with an ever-growing interest in the possible applications it could have for the field. Many questions are being raised as to the possible future realities AI could pose. Out of all these inquiries, one stands out as being particularly dire, a true homage to the ending of a design era, one that as a practice has lasted for hundreds of years and will surely be cut short by the insertion of AI design tools. While this is a melodramatic, foreordained point of view held by a select group, this statement still bears some truth in this figurative death sentence, invoking the question, Can there be a future reality in which artificial intelligence is implemented within the discipline of architecture to develop a symbiosis of the technological and the human aspect, or will AI surpass the need for designers, deeming architects an artifact of a now irrelevant past?
It is first necessary to understand what implications these types of programs can have on a practice like architecture. The process of design in many ways stays relatively the same throughout all disciplines: there is the initial problem, and what follows are the proposed solutions, often falling into a single iterative artifact, all pertaining to a proposed solution to that first problem. This is typically the job of the designer, someone who uses their skills and broad knowledge of the field to efficiently and effectively solve the problem; however, this methodology can be collected over time and eventually mapped out, essentially giving the designer a possible solution without even the slightest physical design or personal iteration. This hints at a reality where the human aspect of the design process becomes a less substantial factor in the development of the project. In Duo Dickinson’s article “Will Architecture in the Future be a Luxury Service?”, a point similar to this is made with reference to the exponential speed in which services are required in today's markets: “People still love growing food and making buildings. Faster, cheaper, easier solutions, via technology, may minimize the human hand in making things in all aspects of our culture.” Using comparisons from other disciplines, like food production, Dickinson highlights the broad nature of this spreading phenomenon and the possible outcomes that could affect more hands-on, repetitive practices. The article further illustrates the shift in design-related technologies’ evolution toward a possible architecture-less discipline: “In architecture, new technologies will inevitably mean that building consumers and contractors will have more ability to bypass architects entirely. Of course, some clients have always prescribed a design and the architects responded to their directives. But now, the next generation of software will respond directly, with even less need for an architect’s input. Data, once coordinated by architects, will be made seamlessly accessible and useful without the cost, pretense and liability of humans, especially those with a design-based education.” This is a possible reality for architects, one that can already be witnessed with small-scale, iterative design processes within the profession. This is further amplified when you take into account the level at which the majority of architects design. Small-scale residential architecture would be the first of the design levels to be affected by this shift in computing advancement, with AI being able to contemplate a multiplicity of simple design iterations in a matter of days--any human intervention would seem irrelevant in terms of both economics and efficiency.
This is only a single reality for the immersion of AI into the discipline of architectural design. There are in fact multiple paths forward in this evolution of design, many of which have a stronger constitution between the human and technological components. There are a variety of different opinions on the burgeoning topic of deep-learning computational machines, all of which hint at, if not shout out, the seemingly inevitable future of AI in replacing the human aspect in the digital and physical world. While there are negatives to this proposed reality, there are in fact working models of human machine symbiosis already happening today. In the article, "I Don't Really See AI as A Threat: Imdat As on Artificial Intelligence in Architecture,” the author, Michael J. Crosbie, discusses the many positive interactions that designers can have with artificially intelligent workflow machines: “With new AI-driven software, a designer might provide a host of constraints, say, a chair made out of a particular material that can hold 300 pounds. The software could generate hundreds of optimal chair variations that the designer could choose from and develop further.” This acts as a perfect example of the AI processing a design problem and, instead of taking the project away from the designer, expediting the initial iterative process, thus saving valuable time and money for both parties. This type of application can be used throughout the field, and being coupled through pre-existing design tools, like Autodesk Suite or Rhino modeling software, could greatly improve output performance of all design projects. This process can be derived simply from deep-learning machines’ ability to catalog data, most of which has already been stored on cloud-based platforms now made available to software companies that are building information databases for their AI projects. This allows the system to compare millions of possible examples and outcomes for a specific design question and problem using artificial levels of computational intuition, a topic that is also noted in Crosbie’s piece: “ Deep-learning machines could decipher patterns in architectural design that architects have intuitively or intentionally created over the years.” Crosbie also states, “AI could potentially expand such patterns to include not only functional and programmatic concerns, but also socio-economic, ideological, geographic, climatic, or other patterns that shape the built environment. With such a resource, AI-software could assemble the best patterns for a given problem into new compositions.” These new, relatively boundless parameters give architects the tools necessary to tackle even bigger problems than before, all while saving the necessary resources to do so.
In this reality, architects have limitless access to centuries of design practice, with a newfound precision in the development and perfection of dynamic, culturally aware social architecture in tandem with an expedited workflow, thus saving time and resources, all thanks to a symbiotic relationship with new AI-based design tools. Even with this template being possible in the near future, AI will inevitably evolve past this point as programs for computational learning surpass the need for any and all human intervention. This ideology spawns from the basic need to advance and improve exciting technologies. Even in disciplines like architecture, the physical ways in which structures are produced have been changing since the very first shelters were being erected thousands of years ago. This, in conjunction with a multiplicity of other evolving relative contexts, such as population growth, climate control, and famine, means that architecture needs to keep moving forward in the built environment as well as in the computational environment to meet projected future demands. In the TED talk “The incredible inventions of intuitive AI,” the presenter, Maurice Conti, explains the very real prospect that, with a very small level of human interaction, AI systems can design and now build more complex objects than ever was thought possible, using an example of a bridge being constructed over a canal in Amsterdam, by artist Joris Laarman and his fabrication firm MX3D. The goal was “to generativity design and robotically print the world's first autonomously manufactured bridge.” Once the team finishes designing the bridge with an AI system, robotic 3D printers will construct the bridge in its entirety--no human interaction required. This type of design fabrication process will be increasingly necessary with growing demands for faster, stronger, cheaper architecture, and, with AI development not showing any signs of slowing in the considerable future, we are sure to see a steady evolution in these self-learning/building systems.
All of these proposed realities place the architect in a world consumed by AI-entangled workflow and mechanical construction methodologies, all of which limit the architect’s ability to function as a designer or possibly even exist within a discipline. As AI grows within the discipline, the true potential of this computational tool will eventually be realized. Since AI is, relatively speaking, in its infancy, this level of pessimism is naturally expected among concerning parties, while, in truth, the development of AI software is still very much within our control. The TED talk, “Can we build AI without losing control over it?”, presented by Sam Harris, speaks to this level of understanding, stating that if software like the ones proposed here are developed in a sensitive and precise manner, a symbiotic relationship can, in fact, develop between AI and humans, in multiple disciplines, such as architecture and other design/construction practices. Articulation of our intentions will be critical for the evolution of artificial systems. In our usual hurried nature to evolve faster, we forget that we must also remember to evolve smarter, not cutting too many corners and inevitably weakening human involvement throughout the built environment. This would act as the final proposed reality of AI integrated architectural design--one in which we still need to work toward a more computationally humanistic system, where both parties can design, evolve, and prosper together, with AI acting as our extended consciousness in the solving of our world’s very real, current conditions and problems.
SOURCES
Dickinson, Duo. "Will Architecture in the Future Be a Luxury Service?" ArchDaily. August 02, 2018. Accessed September 17, 2018. https://www.archdaily.com/899483/will-architecture-in-the-future-be-a-luxury-service.
Crosbie, Michael J. ""I Don't Really See AI as A Threat": Imdat As on Artificial Intelligence in Architecture." ArchDaily. September 30, 2018. Accessed October 24, 2018. https://www.archdaily.com/902978/i-dont-really-see-ai-as-a-threat-imdat-as-on-artificial-intelligence-in-architecture.
Harris, Sam. "Can We Build AI without Losing Control over It?" TED: Ideas worth Spreading. June 2016. Accessed October 24, 2018. https://www.ted.com/talks/sam_harris_can_we_build_ai_without_losing_control_over_it#t-855839.
Conti, Maurice. "The Incredible Inventions of Intuitive AI." TED: Ideas worth Spreading. April 2016. Accessed October 24, 2018. https://www.ted.com/talks/maurice_conti_the_incredible_inventions_of_intuitive_ai?language=en.
The overall presentation was the first thing that caught my eye. The modern design really makes this inviting for the reader to read, and is helped by the images for overall breakup of words, as well as understanding. AI is such a prevalent topic, your knowledge of architecture was apparent here and when it was applied to this topic it aided in the communication of your point of view. The structure of this assignment was great and the topic extremely interesting. Great job.
ReplyDelete-Cole
Kyle,
ReplyDeleteI don't know the first thing about architecture, but I was able to understand the broader point you were making. That is what I really liked about your piece, although it was directly about the relationship between architecture and AI, the basic ideas could have been applied to the combination of AI and any number of disciplines including financial planning and medicine. One suggestion that I would make is that you try to incorporate more rich forms of media. By this I mean video and insightful graphics that help tell your story, not just break up blocks of text. I know you referenced several articles and a TED Talk but inserting pieces of these sources into the body of your text would help give the reader a more seamless experience.
Peter
I really like you work and the enthusiasm you bring in the love of what is a project for a class, but you is a life and everything. I find it energizing and encouraging keep it up. The work look great and your writing is perfect.
ReplyDelete