The Post-Singularity Architect
Introduction
In the following critical analysis, the works of various theorists and researchers in the fields of architecture, design, and computer science are studied and presented to make a case for what will be understood as the “architectural singularity.” As will be illustrated in the work of Raymond Kurzweil, there is increasing evidence pointing towards an inevitable technological “singularity;” a point at which artificial intelligence (AI) may surpass biological intelligence. As our society becomes more technologically advanced, the traditional role that architecture plays will be challenged, and the way that architects design and interact with the built context will shift drastically. Architecture is likely to take on new forms, previously unimaginable, as it nears closer to the architectural singularity. Through this analysis, it is proposed that architects should prepare to find new roles within the relationships between humans, machine intelligence, and the built form.
Machine intelligence already plays a key role in the ways that we consume information and interact with each other and our environment today. In the segments to follow, concepts and questions are proposed to break down and shed some light on where the architecture profession may be headed in the age of the singularity. Some material being analyzed, among other works, includes Ray Kurzweil's book The Singularity is Near: When Humans Transcend Biology, Benjamin Bratton's book The Stack: On Software and Sovereignty, Carl Frey and Michael Osborne's research The Future of Employment: How Susceptible are Jobs to Computerisation?, Margaret Boden's Creativity and Artificial Intelligence, and Tim McGinley's research A Morphogenic Architecture for Intelligent Buildings. This analysis will be broken down into eight discrete segments discussing the premise behind the architectural singularity, the challenges it raises, and how to approach the future of the architectural practice with this speculative outlook in mind.
In the following critical analysis, the works of various theorists and researchers in the fields of architecture, design, and computer science are studied and presented to make a case for what will be understood as the “architectural singularity.” As will be illustrated in the work of Raymond Kurzweil, there is increasing evidence pointing towards an inevitable technological “singularity;” a point at which artificial intelligence (AI) may surpass biological intelligence. As our society becomes more technologically advanced, the traditional role that architecture plays will be challenged, and the way that architects design and interact with the built context will shift drastically. Architecture is likely to take on new forms, previously unimaginable, as it nears closer to the architectural singularity. Through this analysis, it is proposed that architects should prepare to find new roles within the relationships between humans, machine intelligence, and the built form.
Machine intelligence already plays a key role in the ways that we consume information and interact with each other and our environment today. In the segments to follow, concepts and questions are proposed to break down and shed some light on where the architecture profession may be headed in the age of the singularity. Some material being analyzed, among other works, includes Ray Kurzweil's book The Singularity is Near: When Humans Transcend Biology, Benjamin Bratton's book The Stack: On Software and Sovereignty, Carl Frey and Michael Osborne's research The Future of Employment: How Susceptible are Jobs to Computerisation?, Margaret Boden's Creativity and Artificial Intelligence, and Tim McGinley's research A Morphogenic Architecture for Intelligent Buildings. This analysis will be broken down into eight discrete segments discussing the premise behind the architectural singularity, the challenges it raises, and how to approach the future of the architectural practice with this speculative outlook in mind.
The Architectural Singularity
To begin to unpack this topic, it is important to first understand the premise that we may be approaching a technological and subsequently architectural singularity. In the field of physical science, the singularity represents a point at which space and time “merge indistinguishably and cease to have any independent meaning”1 as described by Matt Williams in Universe Today. The use of the term “singularity,” in reference to technological advancement is often attributed to Ray Kurzweil, an American inventor and futurist, and author of books titled: How to Create a Mind, The Age of Spiritual Machines, and The Singularity is Near.
Kurzweil proposes in his book that within the next few decades, artificial intelligence will eclipse human biological intelligence and that by 2045, the singularity will have arrived, and we cannot possibly predict what the world will look like beyond that point.2 Kurzweil is quoted in The Singularity is Near as saying “human-created technology is accelerating and its powers are expanding at an exponential pace.”3
With this statement, Kurzweil is proposing that technology is advancing at a rate that will completely alter the human environment as we know it.
In his book, The Technological Singularity, Murray Shanahan describes how a technological singularity would change human life as we know it, and that current economic and governmental models would be completely altered in the face of new advancements in artificial intelligence and neurotechnology.4 These advancements in artificial intelligence would include the ability to improve performance in an exponential fashion, similar to how the advancement of transistors and integrated circuits followed an exponential curve in processing power throughout the late twentieth century.5 In his work, Shanahan proposes that following this logic, “intelligence itself…would become subject to the law of accelerating returns, and from here to a technological singularity is but a small leap of faith.”6 In other words, the technological singularity could represent a time at which intelligence itself, whether it be artificial or some form of bio-technical hybrid, could grow in processing power at exponential rates, with significant implications on the built world and future growth of human society.
Extending this logic to the architectural realm, the architectural singularity can be understood as a point beyond which we cannot even imagine what the built world of tomorrow will look like. Tim McGinley poses, in A Morphogenic Architecture for Intelligent Buildings, that beyond the singularity architecture may become more like a living organism that, through its design and components, is more intelligent than both its user and designer.7 In order to understand how intelligent buildings function or might come to be within the age of the singularity, it is important first to clearly define and understand the difference between biological and artificial intelligence.
To begin to unpack this topic, it is important to first understand the premise that we may be approaching a technological and subsequently architectural singularity. In the field of physical science, the singularity represents a point at which space and time “merge indistinguishably and cease to have any independent meaning”1 as described by Matt Williams in Universe Today. The use of the term “singularity,” in reference to technological advancement is often attributed to Ray Kurzweil, an American inventor and futurist, and author of books titled: How to Create a Mind, The Age of Spiritual Machines, and The Singularity is Near.
Kurzweil proposes in his book that within the next few decades, artificial intelligence will eclipse human biological intelligence and that by 2045, the singularity will have arrived, and we cannot possibly predict what the world will look like beyond that point.2 Kurzweil is quoted in The Singularity is Near as saying “human-created technology is accelerating and its powers are expanding at an exponential pace.”3
With this statement, Kurzweil is proposing that technology is advancing at a rate that will completely alter the human environment as we know it.
In his book, The Technological Singularity, Murray Shanahan describes how a technological singularity would change human life as we know it, and that current economic and governmental models would be completely altered in the face of new advancements in artificial intelligence and neurotechnology.4 These advancements in artificial intelligence would include the ability to improve performance in an exponential fashion, similar to how the advancement of transistors and integrated circuits followed an exponential curve in processing power throughout the late twentieth century.5 In his work, Shanahan proposes that following this logic, “intelligence itself…would become subject to the law of accelerating returns, and from here to a technological singularity is but a small leap of faith.”6 In other words, the technological singularity could represent a time at which intelligence itself, whether it be artificial or some form of bio-technical hybrid, could grow in processing power at exponential rates, with significant implications on the built world and future growth of human society.
Extending this logic to the architectural realm, the architectural singularity can be understood as a point beyond which we cannot even imagine what the built world of tomorrow will look like. Tim McGinley poses, in A Morphogenic Architecture for Intelligent Buildings, that beyond the singularity architecture may become more like a living organism that, through its design and components, is more intelligent than both its user and designer.7 In order to understand how intelligent buildings function or might come to be within the age of the singularity, it is important first to clearly define and understand the difference between biological and artificial intelligence.
Defining Intelligence
A highly simplified breakdown of biological intelligence is outlined in the chapter titled “Whole Brain Emulation,” in Shanahan’s book, The Technological Singularity. It is illustrated that biological intelligence is made up of neurons, which are a combination of axons and dendrites. Axons can be understood as a neuron’s output, while dendrites can be understood as its input. Where the input section of one neuron is located closely to the output section of another neuron, a synapse can form, essentially creating a link between the two neurons, allowing information flow between them.8 Neurons form a complex network that allows for the communication and processing of information. It is this network of inputs, outputs, and connections that is often the goal to be recreated in artificial intelligence, essentially mimicking the function of a biological brain.
While it may be possible to recreate this communication network artificially using digital technology, this does not guarantee the creation of an “intelligence.” As described by Shanahan, a unique and important property of the biological brain is its plasticity.9 The network of connections between neurons in the brain is under constant change and adaptation throughout the life of the organism. It is this constant change or reworking of neural connections that enables learning and memory in biological intelligence.10 These key properties are a goal of replication in the creation of an artificial intelligence. The creation of a “plastic” machine intelligence, or one that is capable of learning, memory, and informed reconfiguration, brings artificial intelligence significantly closer in function to biological intelligence. The premise of the architectural singularity is as follows: within the architectural realm, artificial intelligences will be capable of performing design tasks to the same level of quality and understanding as a human architect, and intelligent buildings will exist that are as intelligent (if not more) as both their users and designers. The second point of this premise will be the focus of the following section, outlining how exactly a building can be designed and understood as “intelligent,” and what implications this has on the architectural realm.
A highly simplified breakdown of biological intelligence is outlined in the chapter titled “Whole Brain Emulation,” in Shanahan’s book, The Technological Singularity. It is illustrated that biological intelligence is made up of neurons, which are a combination of axons and dendrites. Axons can be understood as a neuron’s output, while dendrites can be understood as its input. Where the input section of one neuron is located closely to the output section of another neuron, a synapse can form, essentially creating a link between the two neurons, allowing information flow between them.8 Neurons form a complex network that allows for the communication and processing of information. It is this network of inputs, outputs, and connections that is often the goal to be recreated in artificial intelligence, essentially mimicking the function of a biological brain.
While it may be possible to recreate this communication network artificially using digital technology, this does not guarantee the creation of an “intelligence.” As described by Shanahan, a unique and important property of the biological brain is its plasticity.9 The network of connections between neurons in the brain is under constant change and adaptation throughout the life of the organism. It is this constant change or reworking of neural connections that enables learning and memory in biological intelligence.10 These key properties are a goal of replication in the creation of an artificial intelligence. The creation of a “plastic” machine intelligence, or one that is capable of learning, memory, and informed reconfiguration, brings artificial intelligence significantly closer in function to biological intelligence. The premise of the architectural singularity is as follows: within the architectural realm, artificial intelligences will be capable of performing design tasks to the same level of quality and understanding as a human architect, and intelligent buildings will exist that are as intelligent (if not more) as both their users and designers. The second point of this premise will be the focus of the following section, outlining how exactly a building can be designed and understood as “intelligent,” and what implications this has on the architectural realm.
Intelligent Buildings
Today, intelligence is used to describe buildings which are designed with lifespan, construction, and management phases in mind, as well as implementation of measures ensuring the building remains adaptable over its entire lifespan, as described by Victor Callaghan in his chapter titled “Intelligent Environments” in Derek Clements-Croome’s book, Intelligent Buildings.11
Intelligent buildings in the age of the singularity will likely take on a different form of intelligence than that used in architectural discourse today. As described in Manic et al.’s work, Intelligent Buildings of the Future, intelligent buildings are composed of networks of smart hubs, sensors, meters, renewable energy generators, and energy storage systems.12 These complex systems within a building produce increasingly large data sets which will require “automated and adaptive approaches to information processing and real-time decision making.”13
Today, there are several metrics designed to measure the “intelligence” of a building, typically in relation to sustainability and building performance. These metrics include BREEAM (Building Research Establishment Environmental Assessment Method), LEED (Leadership in Energy and Environmental Design), LBC (Living Building Challenge), and more. In respect to these rating systems, “intelligent buildings” perform optimally according to performance standards, and automatically respond to user requirements and changes in the building environment.14 While this understanding of building intelligence touches on forward thinking concepts, the intelligent buildings of the future will be capable of processes typically undertaken by humans, as we are capable of “reasoning, planning and learning.”15 This is to say that intelligent buildings will be able to analyze the immense data sets created by their integrated monitoring systems and make informed decisions on adjustments or adaptations needed for improved overall building performance and environmental comfort.
This concept can be further expanded to a larger scale through an understanding of how intelligent systems will interact in the future. To better understand the implications of a global system of intelligence, one can turn to Benjamin Bratton's work, The Stack, where he breaks down the various layers in which software and technology shape and control our lives and the built environment. Bratton proposes that all the world's computation systems form The Stack, a "megastructure" comprised of six layers including Earth, Cloud, City, Address, Interface, and User.16
The Stack illustrates an understanding of the world's various computational systems not as isolated, individual systems, but as a deeply intertwined network of human connection, social interaction, and control.17 In the realm of an architectural singularity, Bratton's Stack can be understood to be an important model within which to work. Viewing our approach to the practice of architecture going forward within the context of an intelligent global computation system will allow us to pose important questions about where the profession is headed and where we as humans can fit into the emerging new social-political system.
The aforementioned definition of intelligent buildings is not all too farfetched from the reality we currently inhabit today. One might argue that there are likely automated systems performing these exact functions already. The turning point at which one could argue for the emergence of the architectural singularity is, as described by McGinley, “the point at which the intelligence of the building’s fabric and systems become superior to the intelligence of the occupants."18
To understand how this might occur, it is important now to define the ways in which machine intelligence might surpasses biological human intelligence. In his work, Shanahan argues that the successful emulation of a human brain marks the breakthrough necessary to pave the way for what he refers to as a “superintelligence.” Shanahan makes two critical arguments regarding brain emulation that support the theory of an architectural singularity. First, if we successfully emulate the function of a human brain through a digital medium, this digital brain can now be copied a near-infinite number of times, only bound by the limits of digital media storage. This marks the first milestone in creating a superintelligence, as it forms the basis for a network or collective intelligence.19 Secondly, Shanahan argues that a digital intelligence is no longer bound to physical constraints, allowing it to be sped up. This combination of collective intelligence performing actions at rates faster than biologically possible forms the basis for Shanahan’s definition of superintelligence and marks a turning point for all human activity.20 With an influx in super-intelligent computation, it will be crucial to understand possible implications on both the architectural practice, and human employment in general.
Today, intelligence is used to describe buildings which are designed with lifespan, construction, and management phases in mind, as well as implementation of measures ensuring the building remains adaptable over its entire lifespan, as described by Victor Callaghan in his chapter titled “Intelligent Environments” in Derek Clements-Croome’s book, Intelligent Buildings.11
Intelligent buildings in the age of the singularity will likely take on a different form of intelligence than that used in architectural discourse today. As described in Manic et al.’s work, Intelligent Buildings of the Future, intelligent buildings are composed of networks of smart hubs, sensors, meters, renewable energy generators, and energy storage systems.12 These complex systems within a building produce increasingly large data sets which will require “automated and adaptive approaches to information processing and real-time decision making.”13
Today, there are several metrics designed to measure the “intelligence” of a building, typically in relation to sustainability and building performance. These metrics include BREEAM (Building Research Establishment Environmental Assessment Method), LEED (Leadership in Energy and Environmental Design), LBC (Living Building Challenge), and more. In respect to these rating systems, “intelligent buildings” perform optimally according to performance standards, and automatically respond to user requirements and changes in the building environment.14 While this understanding of building intelligence touches on forward thinking concepts, the intelligent buildings of the future will be capable of processes typically undertaken by humans, as we are capable of “reasoning, planning and learning.”15 This is to say that intelligent buildings will be able to analyze the immense data sets created by their integrated monitoring systems and make informed decisions on adjustments or adaptations needed for improved overall building performance and environmental comfort.
This concept can be further expanded to a larger scale through an understanding of how intelligent systems will interact in the future. To better understand the implications of a global system of intelligence, one can turn to Benjamin Bratton's work, The Stack, where he breaks down the various layers in which software and technology shape and control our lives and the built environment. Bratton proposes that all the world's computation systems form The Stack, a "megastructure" comprised of six layers including Earth, Cloud, City, Address, Interface, and User.16
The Stack illustrates an understanding of the world's various computational systems not as isolated, individual systems, but as a deeply intertwined network of human connection, social interaction, and control.17 In the realm of an architectural singularity, Bratton's Stack can be understood to be an important model within which to work. Viewing our approach to the practice of architecture going forward within the context of an intelligent global computation system will allow us to pose important questions about where the profession is headed and where we as humans can fit into the emerging new social-political system.
The aforementioned definition of intelligent buildings is not all too farfetched from the reality we currently inhabit today. One might argue that there are likely automated systems performing these exact functions already. The turning point at which one could argue for the emergence of the architectural singularity is, as described by McGinley, “the point at which the intelligence of the building’s fabric and systems become superior to the intelligence of the occupants."18
To understand how this might occur, it is important now to define the ways in which machine intelligence might surpasses biological human intelligence. In his work, Shanahan argues that the successful emulation of a human brain marks the breakthrough necessary to pave the way for what he refers to as a “superintelligence.” Shanahan makes two critical arguments regarding brain emulation that support the theory of an architectural singularity. First, if we successfully emulate the function of a human brain through a digital medium, this digital brain can now be copied a near-infinite number of times, only bound by the limits of digital media storage. This marks the first milestone in creating a superintelligence, as it forms the basis for a network or collective intelligence.19 Secondly, Shanahan argues that a digital intelligence is no longer bound to physical constraints, allowing it to be sped up. This combination of collective intelligence performing actions at rates faster than biologically possible forms the basis for Shanahan’s definition of superintelligence and marks a turning point for all human activity.20 With an influx in super-intelligent computation, it will be crucial to understand possible implications on both the architectural practice, and human employment in general.
Computation and Work
As the argument for an architectural singularity is highly speculative, it is nearly impossible to definitively determine the effects that superintelligence will have on the human workforce. While it already proves difficult to speculate what the workforce would even consist of in the age of the singularity, one can extend current discussion to this line of thinking. To understand where employment is headed in the age of increasingly powerful computation, one can look to Carl Frey and Michael Osborne's 2013 research titled The Future of Employment: How Susceptible are Jobs to Computerisation?
In this work, Frey and Osborne analyze the likelihood of 702 various occupations being rendered obsolete through computerisation in the not-too-distant future. They propose that an estimated 47 percent of jobs in the US are at risk of computerisation, further explaining that we are seeing a polarization of the workforce towards either high-income cognitive jobs, or low-income manual jobs, and away from middle-income routine jobs.21 Frey & Osborne do address that not all jobs in the future would be rendered obsolete solely due to computerisation. They also provide the example of "offshoring,” like we have seen in professions such as customer service or telemarketing. A job is considered “unable to be offshored" if it meets two criteria: the work is location-specific and requires face-to-face personal communication.22 To demonstrate this, Frey and Osborne present the example of cashiers that have largely been replaced with self-serve technology. This task must both be performed at a specific location and requires face-to-face contact. Therefore, they cannot be offshored, but it can be automated with relative ease.23
To support the notion that human architects are well poised to remain relevant in the workforce despite rapid computerisation, Frey and Osborne place architects as the 82nd least likely profession to be computerised out of a total 702 occupations. Frey and Osborne’s methodology consists of ranking each individual profession by probability of computerisation, according to nine discreet variables. These variables are as follows: Assisting and caring for others, Persuasion, Negotiation, Social Perceptiveness, Fine arts, Originality, Manual dexterity, Finger dexterity, and Cramped workspace.24 Some might argue Frey and Osborne’s classification of architects as 82nd least likely to be computerized to be inaccurate due to the difficulty of truly grasping the job function of an architect without experiencing the role first-hand. If one regards this classification generally, there is still a strong indication that architects will continue to be relevant, in one form or another, despite increasing computerization.
As the argument for an architectural singularity is highly speculative, it is nearly impossible to definitively determine the effects that superintelligence will have on the human workforce. While it already proves difficult to speculate what the workforce would even consist of in the age of the singularity, one can extend current discussion to this line of thinking. To understand where employment is headed in the age of increasingly powerful computation, one can look to Carl Frey and Michael Osborne's 2013 research titled The Future of Employment: How Susceptible are Jobs to Computerisation?
In this work, Frey and Osborne analyze the likelihood of 702 various occupations being rendered obsolete through computerisation in the not-too-distant future. They propose that an estimated 47 percent of jobs in the US are at risk of computerisation, further explaining that we are seeing a polarization of the workforce towards either high-income cognitive jobs, or low-income manual jobs, and away from middle-income routine jobs.21 Frey & Osborne do address that not all jobs in the future would be rendered obsolete solely due to computerisation. They also provide the example of "offshoring,” like we have seen in professions such as customer service or telemarketing. A job is considered “unable to be offshored" if it meets two criteria: the work is location-specific and requires face-to-face personal communication.22 To demonstrate this, Frey and Osborne present the example of cashiers that have largely been replaced with self-serve technology. This task must both be performed at a specific location and requires face-to-face contact. Therefore, they cannot be offshored, but it can be automated with relative ease.23
To support the notion that human architects are well poised to remain relevant in the workforce despite rapid computerisation, Frey and Osborne place architects as the 82nd least likely profession to be computerised out of a total 702 occupations. Frey and Osborne’s methodology consists of ranking each individual profession by probability of computerisation, according to nine discreet variables. These variables are as follows: Assisting and caring for others, Persuasion, Negotiation, Social Perceptiveness, Fine arts, Originality, Manual dexterity, Finger dexterity, and Cramped workspace.24 Some might argue Frey and Osborne’s classification of architects as 82nd least likely to be computerized to be inaccurate due to the difficulty of truly grasping the job function of an architect without experiencing the role first-hand. If one regards this classification generally, there is still a strong indication that architects will continue to be relevant, in one form or another, despite increasing computerization.
To provide some clarity regarding their rankings, Frey and Osborne outline some bottlenecks of computerisation of various industries. These are aspects of the work that make a particular profession challenging to automate through computerisation, and include Perception and Manipulation, Social Intelligence, and Creative Intelligence.25 It will be increasingly important to address the latter two aspects outlined as bottlenecks by Frey and Osborne, as they will help to inform the role of the post-singularity architect going forward.
It is proposed in this critical analysis that while the “architect” may continue to exist despite increased computational power, the definition of this role is likely to undergo a significant transformation as it approaches the architectural singularity.
Artificial Creativity
To address the concept of creativity in relation to artificial intelligence, one can turn to Margaret Boden's research article titled Creativity and Artificial Intelligence. Boden's work proposes that artificial intelligence can prove useful in creative work in three ways:
A key takeaway from Boden's work is that AI will likely be much better at creating or modelling new ideas than it will be at evaluating their success in the lived human context. Boden describes that the challenges facing an AI regarding creative thinking lie within the concept of “domain-expertise” and “valuation.”27 This is to say that newly implemented AI systems lack experience within the domain or realm which they are inserted into. Like a human intern starting work in an established firm, the AI may have been “trained” with traditional knowledge in building science or architectural design, but it lacks the ability to evaluate its performance without feedback from an overseeing entity. This limitation of artificial intelligence is outlined in Martin Rooney and Steven Smith’s research titled Artificial Intelligence in Engineering Design.
In their work, Rooney and Smith define the difference between objective and subjective cases within the realm of design. Objective cases refer to aspects of design that can be addressed through calculations and can be applied consistently between projects.28 Subjective cases in this realm refer to aspects of the design process that traditionally require human input, as they typically deal with aesthetic decisions. Subjective cases can vary from one design to another and are rarely able to be widely applied like objective design aspects.29 In order for artificial intelligence to be able to evaluate its performance in subjective design decisions, it requires a feedback mechanism. The feedback mechanism outlined by Rooney and Smith works to apply previous design experience to a problem, extract relevant information to be used in a subsequent problem, and store the extracted information for later use.30 As described by Rooney and Smith, today’s human designer typically carries out these feedback steps subconsciously as they progress through the design process. Outlined in this research is the key to one future role of the post-singularity architect, where it is stated that an intelligent “program” would be able to change its decisions or “self-modify” according to the previous successes and failures outlined in the aforementioned feedback loop.31
It is proposed in this critical analysis that while the “architect” may continue to exist despite increased computational power, the definition of this role is likely to undergo a significant transformation as it approaches the architectural singularity.
Artificial Creativity
To address the concept of creativity in relation to artificial intelligence, one can turn to Margaret Boden's research article titled Creativity and Artificial Intelligence. Boden's work proposes that artificial intelligence can prove useful in creative work in three ways:
- It can produce novel combinations of familiar ideas
-
It can explore the potential of conceptual spaces
-
It can make transformations that enable the generation of previously impossible ideas26
A key takeaway from Boden's work is that AI will likely be much better at creating or modelling new ideas than it will be at evaluating their success in the lived human context. Boden describes that the challenges facing an AI regarding creative thinking lie within the concept of “domain-expertise” and “valuation.”27 This is to say that newly implemented AI systems lack experience within the domain or realm which they are inserted into. Like a human intern starting work in an established firm, the AI may have been “trained” with traditional knowledge in building science or architectural design, but it lacks the ability to evaluate its performance without feedback from an overseeing entity. This limitation of artificial intelligence is outlined in Martin Rooney and Steven Smith’s research titled Artificial Intelligence in Engineering Design.
In their work, Rooney and Smith define the difference between objective and subjective cases within the realm of design. Objective cases refer to aspects of design that can be addressed through calculations and can be applied consistently between projects.28 Subjective cases in this realm refer to aspects of the design process that traditionally require human input, as they typically deal with aesthetic decisions. Subjective cases can vary from one design to another and are rarely able to be widely applied like objective design aspects.29 In order for artificial intelligence to be able to evaluate its performance in subjective design decisions, it requires a feedback mechanism. The feedback mechanism outlined by Rooney and Smith works to apply previous design experience to a problem, extract relevant information to be used in a subsequent problem, and store the extracted information for later use.30 As described by Rooney and Smith, today’s human designer typically carries out these feedback steps subconsciously as they progress through the design process. Outlined in this research is the key to one future role of the post-singularity architect, where it is stated that an intelligent “program” would be able to change its decisions or “self-modify” according to the previous successes and failures outlined in the aforementioned feedback loop.31
Conclusion
To summarize the findings of this analysis, three scenarios are proposed as the likely future role of the architect in the coming age of the architectural singularity: Architect as Teacher, Architect as Curator, and Architect as Luxury. These scenarios could occur individually or simultaneously, forming a new hybrid model for the architect. These models could also form the new major career paths in the architectural field. Like how many students today studying architecture in tertiary education are presented with the choice between architecture, building science, and project management, the practice in the age of the architectural singularity may morph into a system that prepares students to enter the practice as evaluator or curator of AI-designed architecture, or as “legacy architect,” a service only accessible to the hyper-rich.
There will be a need for someone to define the constraints within which the AI is programmed to work. Like setting site boundaries or a maximum height in CAD and BIM software, the AI needs to know which parameters to consider and which to ignore when ideating new designs or when working towards a new “optimal” form for a given context.
Secondly, as mentioned in Boden's work, AIs may not yet be able to properly evaluate their performance without feedback from humans. There will be a need for human evaluation against project requirements and overall current cultural values against which an AI may not be equipped to evaluate. As human culture and values are constantly changing, and since architecture reflects our culture, the AI will need to be constantly fed new information and evaluated on its performance.
Finally, as the price of computation is driven downward by increased adoption of new technologies, there is an increasingly strong case for the possibility of AI that can design architectural solutions more efficiently than their human counterparts. These buildings may however be part of a limited class, at least in the beginning of this new technological era, that are bound by straightforward criteria and end-goals. The section to follow expands on the proposal of these three possible roles for the post-singularity architect.
Architect as Teacher
As noted in Boden’s work, the weakness of AI lies in its ability to self-evaluate. This is where the first role of the post-singularity architect comes into play: the architect as teacher. The architect as teacher acts as a liaison between the values and desires humans posses for their architectural projects, and the computational power of the AI-architect. By programming the specific parameters within which the AI is to design, the architect as teacher influences the design outcome by guiding the AI through which parameters are to be optimized for and which are to be ignored. By defining the constraints within which the AI is intended to design, the architect as teacher prepares the AI-architect to be passed along to the next stage in the new-age design process: evaluation.
Understanding the architect as teacher or evaluator implies some level of authority maintained over machine. The AI-architect could be understood as being given “limited freedom” in the sense that it is free to ideate and test the optimal architectural forms within a confined set of specified parameters.
As noted in Rooney and Smith’s research on artificial intelligence in design, it will be important to maintain human control over artificially intelligent designers, especially when it comes to subjective design decisions. Rooney and Smith are quoted as stating, “it is important for professional liability that the human designer remain in control and the artificial intelligence component act as an experience consultant.”32 In other words, artificial intelligence should be understood as a powerful tool in design, but not as the governing body or overseer. The role of overseer is to be maintained as a human position: the architect as curator. Maintaining some level of human control over the outcome of AI-designed projects will allow us to maintain cultural and aesthetic value in our designs, even beyond the singularity.
To summarize the findings of this analysis, three scenarios are proposed as the likely future role of the architect in the coming age of the architectural singularity: Architect as Teacher, Architect as Curator, and Architect as Luxury. These scenarios could occur individually or simultaneously, forming a new hybrid model for the architect. These models could also form the new major career paths in the architectural field. Like how many students today studying architecture in tertiary education are presented with the choice between architecture, building science, and project management, the practice in the age of the architectural singularity may morph into a system that prepares students to enter the practice as evaluator or curator of AI-designed architecture, or as “legacy architect,” a service only accessible to the hyper-rich.
There will be a need for someone to define the constraints within which the AI is programmed to work. Like setting site boundaries or a maximum height in CAD and BIM software, the AI needs to know which parameters to consider and which to ignore when ideating new designs or when working towards a new “optimal” form for a given context.
Secondly, as mentioned in Boden's work, AIs may not yet be able to properly evaluate their performance without feedback from humans. There will be a need for human evaluation against project requirements and overall current cultural values against which an AI may not be equipped to evaluate. As human culture and values are constantly changing, and since architecture reflects our culture, the AI will need to be constantly fed new information and evaluated on its performance.
Finally, as the price of computation is driven downward by increased adoption of new technologies, there is an increasingly strong case for the possibility of AI that can design architectural solutions more efficiently than their human counterparts. These buildings may however be part of a limited class, at least in the beginning of this new technological era, that are bound by straightforward criteria and end-goals. The section to follow expands on the proposal of these three possible roles for the post-singularity architect.
Architect as Teacher
As noted in Boden’s work, the weakness of AI lies in its ability to self-evaluate. This is where the first role of the post-singularity architect comes into play: the architect as teacher. The architect as teacher acts as a liaison between the values and desires humans posses for their architectural projects, and the computational power of the AI-architect. By programming the specific parameters within which the AI is to design, the architect as teacher influences the design outcome by guiding the AI through which parameters are to be optimized for and which are to be ignored. By defining the constraints within which the AI is intended to design, the architect as teacher prepares the AI-architect to be passed along to the next stage in the new-age design process: evaluation.
Understanding the architect as teacher or evaluator implies some level of authority maintained over machine. The AI-architect could be understood as being given “limited freedom” in the sense that it is free to ideate and test the optimal architectural forms within a confined set of specified parameters.
As noted in Rooney and Smith’s research on artificial intelligence in design, it will be important to maintain human control over artificially intelligent designers, especially when it comes to subjective design decisions. Rooney and Smith are quoted as stating, “it is important for professional liability that the human designer remain in control and the artificial intelligence component act as an experience consultant.”32 In other words, artificial intelligence should be understood as a powerful tool in design, but not as the governing body or overseer. The role of overseer is to be maintained as a human position: the architect as curator. Maintaining some level of human control over the outcome of AI-designed projects will allow us to maintain cultural and aesthetic value in our designs, even beyond the singularity.
Architect as Curator
With the emergence of a new wealth of machine-born architectural designs, a new human role is proposed. While artificial intelligence may be able to be programmed to produce designs under specific conditions, within specific parameters, as is outlined in Boden’s work, artificial intelligence may not yet be able to evaluate the success of these designs. In other words, while there could be an infinite amount of design iterations presented for a particular architectural project, there may be a need for human intervention in the decision-making process for which design best suits the needs of the client.
Tim McGinley discusses the possible future role of the architect as forming a sort of “bridge” between desired architectural solutions and self-organizing systems. For instance, if a client provided specific criteria to be met for a desired building project, and an AI-architect was programmed with these specific parameters to generate a multitude of “ideal” design choices, based on the given criteria, it now becomes the role of the architect as curator to review these options through a critical, human-oriented lens and decide which is most successful for the project brief.
This example highlights the strong relation between the roles of the architect as teacher and curator, as these roles could be understood to form a cyclical relationship. As the architect (as teacher) defines specific parameters within which an AI-architect is to work, the AI learns which variables are considered valuable in various design conditions. As the architect (as curator) chooses or rejects proposed designs, the AI-architect learns which projects were successful or unsuccessful and why or why not. In this sense, the architect as curator acts as a form of direct evaluator, as well as an indirect teacher, constantly giving the AI feedback on proposed design interventions and feeding it information regarding which proposals were most liked by clients.
Following the logic of this feedback loop, one could argue that through this relationship, the human architect (as curator or teacher) could render themself obsolete by providing the AI with compounding data on successful architectural projects. Machine learning depends on large datasets to learn from and only improves in accuracy as more “feedback” is provided. In this sense, it could be argued that the architect, in providing feedback to the AI, could be laying the groundwork for an AI that is eventually knowledgeable enough to make decisions regarding the success of an architectural intervention without the feedback of a human architect. At the very least, the AI could be “intelligent” enough to predict the success of a design to a close degree of accuracy.
This argument might be sound if humanity were a static entity. However, humanity is very much a dynamic, ever-evolving entity. This is where the question of culture comes into play. Culture varies drastically, dependent on many factors including geographic location, population demographics, religion, and more. Traditional or “legacy” architecture is typically reflective of the cultural values of the place and time in which it is built. For the same reason that some criticize globalized architecture produced today by what we know as “starchitects,” an AI could likely never reach a level at which it could objectively decide the most appropriate or successful architectural form for a given context. This is because of humanity’s inherent dynamic nature. As human culture continually evolves, it can be argued that in order for an AI-architect to remain current or “up-to-date,” it will require the architect as teacher to continually provide information regarding cultural and aesthetic value in order to continue to ideate successful and relevant architectural designs.
![Fig. 1 Post-Singularity Architect Process.]()
Figure 1 illustrates the relationship of the architect as teacher and curator in a workflow diagram influenced by Rooney and Smith’s 1983 intelligent CAD model illustrations. In this model, the architect as teacher works to “program” various constraints, values, and parameters to be optimize within the artificially intelligent design program. The “AI-architect” then begins the process of ideation, iteration testing, and analysis informed by the parameters input by the human teacher. The AI-architect then presents the “optimal” design schemes to the architect as curator, who decides on a final revision or restarts the process again with new constraints or variables. Once an acceptable design has been chosen, it can be presented to the client and accepted or reworked using the model outlined above. Once a successful design scheme has been chosen, the AI-architect then extracts relevant information to be stored and used as learning experience for the next design proposal. Within this theoretical model, the AI-architect learns how to better perform with each new application of the process. The human architect as teacher and curator remains, however, to ensure that the AI model is properly trained and to maintain human oversight in the decision process of the final revision.
With the emergence of a new wealth of machine-born architectural designs, a new human role is proposed. While artificial intelligence may be able to be programmed to produce designs under specific conditions, within specific parameters, as is outlined in Boden’s work, artificial intelligence may not yet be able to evaluate the success of these designs. In other words, while there could be an infinite amount of design iterations presented for a particular architectural project, there may be a need for human intervention in the decision-making process for which design best suits the needs of the client.
Tim McGinley discusses the possible future role of the architect as forming a sort of “bridge” between desired architectural solutions and self-organizing systems. For instance, if a client provided specific criteria to be met for a desired building project, and an AI-architect was programmed with these specific parameters to generate a multitude of “ideal” design choices, based on the given criteria, it now becomes the role of the architect as curator to review these options through a critical, human-oriented lens and decide which is most successful for the project brief.
This example highlights the strong relation between the roles of the architect as teacher and curator, as these roles could be understood to form a cyclical relationship. As the architect (as teacher) defines specific parameters within which an AI-architect is to work, the AI learns which variables are considered valuable in various design conditions. As the architect (as curator) chooses or rejects proposed designs, the AI-architect learns which projects were successful or unsuccessful and why or why not. In this sense, the architect as curator acts as a form of direct evaluator, as well as an indirect teacher, constantly giving the AI feedback on proposed design interventions and feeding it information regarding which proposals were most liked by clients.
Following the logic of this feedback loop, one could argue that through this relationship, the human architect (as curator or teacher) could render themself obsolete by providing the AI with compounding data on successful architectural projects. Machine learning depends on large datasets to learn from and only improves in accuracy as more “feedback” is provided. In this sense, it could be argued that the architect, in providing feedback to the AI, could be laying the groundwork for an AI that is eventually knowledgeable enough to make decisions regarding the success of an architectural intervention without the feedback of a human architect. At the very least, the AI could be “intelligent” enough to predict the success of a design to a close degree of accuracy.
This argument might be sound if humanity were a static entity. However, humanity is very much a dynamic, ever-evolving entity. This is where the question of culture comes into play. Culture varies drastically, dependent on many factors including geographic location, population demographics, religion, and more. Traditional or “legacy” architecture is typically reflective of the cultural values of the place and time in which it is built. For the same reason that some criticize globalized architecture produced today by what we know as “starchitects,” an AI could likely never reach a level at which it could objectively decide the most appropriate or successful architectural form for a given context. This is because of humanity’s inherent dynamic nature. As human culture continually evolves, it can be argued that in order for an AI-architect to remain current or “up-to-date,” it will require the architect as teacher to continually provide information regarding cultural and aesthetic value in order to continue to ideate successful and relevant architectural designs.
Figure 1 illustrates the relationship of the architect as teacher and curator in a workflow diagram influenced by Rooney and Smith’s 1983 intelligent CAD model illustrations. In this model, the architect as teacher works to “program” various constraints, values, and parameters to be optimize within the artificially intelligent design program. The “AI-architect” then begins the process of ideation, iteration testing, and analysis informed by the parameters input by the human teacher. The AI-architect then presents the “optimal” design schemes to the architect as curator, who decides on a final revision or restarts the process again with new constraints or variables. Once an acceptable design has been chosen, it can be presented to the client and accepted or reworked using the model outlined above. Once a successful design scheme has been chosen, the AI-architect then extracts relevant information to be stored and used as learning experience for the next design proposal. Within this theoretical model, the AI-architect learns how to better perform with each new application of the process. The human architect as teacher and curator remains, however, to ensure that the AI model is properly trained and to maintain human oversight in the decision process of the final revision.
Architect as Luxury
As machine-designed architecture gains traction due to its financial incentives, architects as we currently know them may become a niche within the practice, referred to in this work as “legacy architects.” With most domestic architecture being designed by AI which churns out generic solutions to static housing typologies, and commercial architecture being outsourced to powerful algorithmic design studios designing for an optimal balance between predetermined criteria, legacy architects become something of a luxury that only the ultra rich can afford.
This dynamic could work similarly to purchasing a built-to-order, hand-made custom Rolls Royce today. With claims that this method of construction is somehow “more authentic” and exclusive, the markup renders such objects only attainable by the ultra rich. This model remains financially viable as the low volume vehicles are sold at exorbitant prices. Architecture, in this sense, could operate as a niche practice of a select few providing a “hand-made” type of service reminiscent of what will be understood as the legacy architecture of today.
It is entirely possible that small to mid to large-size architecture firms will begin to outsource their design efforts to machine intelligence, while ultra-popular “starchitects” will be presented with opportunities to pick and choose specialized projects demanding a “human touch,” hired for outlandish financial compensation.
As machine-designed architecture gains traction due to its financial incentives, architects as we currently know them may become a niche within the practice, referred to in this work as “legacy architects.” With most domestic architecture being designed by AI which churns out generic solutions to static housing typologies, and commercial architecture being outsourced to powerful algorithmic design studios designing for an optimal balance between predetermined criteria, legacy architects become something of a luxury that only the ultra rich can afford.
This dynamic could work similarly to purchasing a built-to-order, hand-made custom Rolls Royce today. With claims that this method of construction is somehow “more authentic” and exclusive, the markup renders such objects only attainable by the ultra rich. This model remains financially viable as the low volume vehicles are sold at exorbitant prices. Architecture, in this sense, could operate as a niche practice of a select few providing a “hand-made” type of service reminiscent of what will be understood as the legacy architecture of today.
It is entirely possible that small to mid to large-size architecture firms will begin to outsource their design efforts to machine intelligence, while ultra-popular “starchitects” will be presented with opportunities to pick and choose specialized projects demanding a “human touch,” hired for outlandish financial compensation.
Some may argue that architecture designed by artificial intelligence is somehow less human, or removed from humanity, in comparison to what we understand today as “traditional,” human-designed architecture. By extension, this argument implies that these “less human” spaces are somehow less desirable than human-designed spaces. To address this perspective, one could argue that, since artificial intelligence is ultimately created by humans, by extension, the things that AI creates can also be counted within the realm of human creation. To further argue this point, if artificial intelligence is tasked with ideating and creating architecture for humankind, the architecture remains deeply connected to humanity through its very purpose, maintaining its definition as an ultimately human endeavour.
There is agreement among researchers and theorists that the singularity is coming, with less agreement on exactly when it may come. Now is a crucial time to consider the structure within which architects will work given these speculative advancements in machine intelligence. If architects and others in the field approach the problem carefully, we may be able to carve out a lasting role even beyond the architectural singularity. This role will depend highly on the development of artificial intelligence and its success within the design industry. Despite the likely implementation of AI into the practice, human architects are likely to remain firmly integrated into society as critical evaluators and curators within the feedback loop enabling the creation of architecture beyond the era of the singularity.
There is agreement among researchers and theorists that the singularity is coming, with less agreement on exactly when it may come. Now is a crucial time to consider the structure within which architects will work given these speculative advancements in machine intelligence. If architects and others in the field approach the problem carefully, we may be able to carve out a lasting role even beyond the architectural singularity. This role will depend highly on the development of artificial intelligence and its success within the design industry. Despite the likely implementation of AI into the practice, human architects are likely to remain firmly integrated into society as critical evaluators and curators within the feedback loop enabling the creation of architecture beyond the era of the singularity.
Footnotes
- Matt Williams, “What Is A Singularity?,” Universe Today, last modified January 7, 2017, https://www.universetoday.com/84147/singularity/.
-
Ray Kurzweil, The Singularity Is Near: When Humans Transcend Biology (London: Duckworth, 2016), 10.
- Kurzweil, 10.
- Murray Shanahan, “Introduction,” in The Technological Singularity (Cambridge: The MIT Press, 2015a), xv.
- Kurzweil, xviii.
- Shanahan, xix.
- Tim McGinley, “A Metamorphogenetic Architecture for Intelligent Buildings,” Intelligent Buildings International 7, no. 1 (2014): 4, https://doi.org/10.1080/17508975.2014.970120
-
Murray Shanahan, “Whole Brain Emulation,” in The Technological Singularity (Cambridge: The MIT Press, 2015b), 15.
- Shanahan, 17.
- Shanahan, 17.
- Victor Callaghan, “Chapter 5: Intelligent Environments.” in Intelligent Buildings (London: ICE Publishing, 2013a), 72, http://dx.doi.org/10.1680/ib.57340.071.
- Milos Manic et al., “Intelligent Buildings of the Future: Cyberaware, Deep Learning Powered, and Human Interacting,” IEEE Industrial Electronics Magazine 10, no. 4 (2016), 38, https://doi.org/10.1109/mie.2016.2615575.
- Milos Manic et al., 38.
- Callaghan, 72.
- Callaghan, 72.
-
Benjamin Bratton, The Stack - On Software and Sovereignty. (Massachusetts: MIT Press, 2016), xviii.
- Bratton, xviii.
- McGinley, 4.
- Shanahan, 36.
- Shanahan, 36.
-
Carl Frey and Michael Osborne, “The Future of Employment: How Susceptible Are Jobs to Computerisation?,” Technological Forecasting and Social Change 114 (2013): 3, https://doi.org/10.1016/j.techfore.2016.08.019.
- Frey and Osborne, 5.
- Frey and Osborne, 5.
- Frey and Osborne, 31.
- Frey and Osborne, 31.
- Margaret Boden, “Creativity and Artificial Intelligence.” Artificial Intelligence 103, no. 1-2 (1998): 347, https://doi.org/10.1016/s0004-3702(98)00055-1.
- Boden, 355.
- Martin Rooney and Steven Smith, “Artificial Intelligence in Engineering Design.” Computers & Structures 16, no. 1 (1983): 281, https://doi.org/10.1016/0045-7949(83)90167-0.
- Rooney and Smith, 281.
- Rooney and Smith, 281.
- Rooney and Smith, 281.
- Rooney and Smith, 282.
Bibliography
“A different dystopia July 2030; If robots don’t take all the jobs.” The Economist, vol. 432, no. 9150, 6 July 2019, p. 9(US). Gale OneFile: CPI.Q, https://link-gale-com.ez-proxy.lib.ryerson.ca/apps/doc/A592162382/CPI?u=rpu_main&sid=CPI&xid=96a0d8e1 . Accessed 24 Sept. 2020.’
Baldwin, Eric. “How Artificial Intelligence Will Shape Design by 2050” 16 Apr 2020. ArchDaily. Accessed 24 Sep 2020. ISSN 0719-8884
Boden, Margaret A. “Creativity and Artificial Intelligence.” Artificial Intelligence 103, no. 1-2 (1998): 347–56. https://doi.org/10.1016/s0004-3702(98)00055-1.
Bratton, Benjamin H. The Stack - On Software and Sovereignty. ProQuest Ebook Central. Massachusetts: MIT Press, 2016. https://ebookcentral-proquest-com.ezproxy.lib.ryerson.ca/lib/ryerson/detail.action?docID=4424500.
Carta, Silvio, Stephanie St. Loe, Tommaso Turchi, and Joel Simon. “Self-Organising Floor Plans in Care Homes.” Sustainability, 4393, 12, no. 11 (2020). https://doi.org/10.3390/su12114393.
Carta, Silvio. “Machine Learning and Computational Design.” Ubiquity 2020, no. May (2020). https://doi.org/10.1145/3401842.
Cevora, George. “The Relationship between Biological and Artificial Intelligence.” Illumr, 2019. https://arxiv.org/abs/1905.00547.
Clements-Croome, Derek, and Victor Callaghan. “Chapter 5: Intelligent Environments.” In Intelligent Buildings, 71–87. London: ICE Publishing, 2013a. http://dx.doi.org/10.1680/ib.57340.071.
Clements-Croome, Derek, and Ziona Strelitz. “Chapter 14: The changing culture of living and working: physical and virtual modalities.” In Intelligent Buildings, 215–222. London: ICE Publishing, 2013b. https://doi.org/10.1680/ib.57340.215.
“A different dystopia July 2030; If robots don’t take all the jobs.” The Economist, vol. 432, no. 9150, 6 July 2019, p. 9(US). Gale OneFile: CPI.Q, https://link-gale-com.ez-proxy.lib.ryerson.ca/apps/doc/A592162382/CPI?u=rpu_main&sid=CPI&xid=96a0d8e1 . Accessed 24 Sept. 2020.’
Baldwin, Eric. “How Artificial Intelligence Will Shape Design by 2050” 16 Apr 2020. ArchDaily. Accessed 24 Sep 2020.
Boden, Margaret A. “Creativity and Artificial Intelligence.” Artificial Intelligence 103, no. 1-2 (1998): 347–56. https://doi.org/10.1016/s0004-3702(98)00055-1.
Bratton, Benjamin H. The Stack - On Software and Sovereignty. ProQuest Ebook Central. Massachusetts: MIT Press, 2016. https://ebookcentral-proquest-com.ezproxy.lib.ryerson.ca/lib/ryerson/detail.action?docID=4424500.
Carta, Silvio, Stephanie St. Loe, Tommaso Turchi, and Joel Simon. “Self-Organising Floor Plans in Care Homes.” Sustainability, 4393, 12, no. 11 (2020). https://doi.org/10.3390/su12114393.
Carta, Silvio. “Machine Learning and Computational Design.” Ubiquity 2020, no. May (2020). https://doi.org/10.1145/3401842.
Cevora, George. “The Relationship between Biological and Artificial Intelligence.” Illumr, 2019. https://arxiv.org/abs/1905.00547.
Clements-Croome, Derek, and Victor Callaghan. “Chapter 5: Intelligent Environments.” In Intelligent Buildings, 71–87. London: ICE Publishing, 2013a. http://dx.doi.org/10.1680/ib.57340.071.
Clements-Croome, Derek, and Ziona Strelitz. “Chapter 14: The changing culture of living and working: physical and virtual modalities.” In Intelligent Buildings, 215–222. London: ICE Publishing, 2013b. https://doi.org/10.1680/ib.57340.215.
Clements-Croome, Derek, Michael Beaven, and Duncan Wilson. “Chapter 16: Sky Studios, Hounslow, UK, and the internet of things.” In Intelligent Buildings, 245–270. London: ICE Publishing, 2013c. https://doi.org/10.1680/ib.57340.245.
Fazi, M Beatrice. “Can a Machine Think (Anything New)? Automation beyond Simulation.” AI & Society, 2019, 34, no. 813-824 (February 12, 2018). https://doi.org/10.1007/s00146-018-0821-0.
Frey, Carl Benedikt, and Michael A. Osborne. “The Future of Employment: How Susceptible Are Jobs to Computerisation?” Technological Forecasting and Social Change 114 (2013): 254–80. https://doi.org/10.1016/j.techfore.2016.08.019.
Kurzweil, Ray. The Singularity Is Near: When Humans Transcend Biology. London: Duckworth, 2016.
Lang, Peter, and William Menking. Superstudio: Life Without Objects. Milano: Skira Editore S.p.A., 2003. https://reserve.library.ryerson.ca/ares/ares.dll?SessionID=Y-021920630J&Action=10&Type=10&Value=336760.
Manic, Milos, Kasun Amarasinghe, Juan J. Rodriguez-Andina, and Craig Rieger. “Intelligent Buildings of the Future: Cyberaware, Deep Learning Powered, and Human Interacting.” IEEE Industrial Electronics Magazine 10, no. 4 (2016): 32–49. https://doi.org/10.1109/mie.2016.2615575.
McGinley, Tim. “A Metamorphogenetic Architecture for Intelligent Buildings.” Intelligent Buildings International 7, no. 1 (November 7, 2014): 4–15. https://doi.org/10.1080/17508975.2014.970120.
Rooney, Martin F, and Steven E Smith. “Artificial Intelligence in Engineering Design.” Computers & Structures 16, no. 1 (1983): 279–88. https://doi.org/10.1016/0045-7949(83)90167-0.
Shanahan, Murray. Introduction. In The Technological Singularity, xv-xxiii. Cambridge: The MIT Press, 2015a. https://ebookcentral-proquest-com.ezproxy.lib.ryerson.ca/lib/ryerson/reader.action?docID=3433796&ppg=1.
Shanahan, Murray. “Whole Brain Emulation.” In The Technological Singularity, 15–50. Cambridge: The MIT Press, 2015b. https://ebookcentral-proquest-com.ezproxy.lib.ryerson.ca/lib/ryerson/reader.action?docID=3433796&ppg=75.
Shanahan, Murray. “Superintelligence.” In The Technological Singularity, 85-116. Cambridge: The MIT Press, 2015c. https://ebookcentral-proquest-com.ezproxy.lib.ryerson.ca/lib/ryerson/reader.action?docID=3433796&ppg=110.
Tamke, Martin, Paul Nicholas, and Mateusz Zwierzycki. “Machine Learning for Architectural Design: Practices and Infrastructure.” International Journal of Architectural Computing 16, no. 2 (2018): 123–43. https://doi.org/10.1177/1478077118778580.
Williams, Matt. “What Is A Singularity?.” Universe Today. January 7, 2017. https://www.universetoday.com/84147/singularity/.
Fazi, M Beatrice. “Can a Machine Think (Anything New)? Automation beyond Simulation.” AI & Society, 2019, 34, no. 813-824 (February 12, 2018). https://doi.org/10.1007/s00146-018-0821-0.
Frey, Carl Benedikt, and Michael A. Osborne. “The Future of Employment: How Susceptible Are Jobs to Computerisation?” Technological Forecasting and Social Change 114 (2013): 254–80. https://doi.org/10.1016/j.techfore.2016.08.019.
Kurzweil, Ray. The Singularity Is Near: When Humans Transcend Biology. London: Duckworth, 2016.
Lang, Peter, and William Menking. Superstudio: Life Without Objects. Milano: Skira Editore S.p.A., 2003. https://reserve.library.ryerson.ca/ares/ares.dll?SessionID=Y-021920630J&Action=10&Type=10&Value=336760.
Manic, Milos, Kasun Amarasinghe, Juan J. Rodriguez-Andina, and Craig Rieger. “Intelligent Buildings of the Future: Cyberaware, Deep Learning Powered, and Human Interacting.” IEEE Industrial Electronics Magazine 10, no. 4 (2016): 32–49. https://doi.org/10.1109/mie.2016.2615575.
McGinley, Tim. “A Metamorphogenetic Architecture for Intelligent Buildings.” Intelligent Buildings International 7, no. 1 (November 7, 2014): 4–15. https://doi.org/10.1080/17508975.2014.970120.
Rooney, Martin F, and Steven E Smith. “Artificial Intelligence in Engineering Design.” Computers & Structures 16, no. 1 (1983): 279–88. https://doi.org/10.1016/0045-7949(83)90167-0.
Shanahan, Murray. Introduction. In The Technological Singularity, xv-xxiii. Cambridge: The MIT Press, 2015a. https://ebookcentral-proquest-com.ezproxy.lib.ryerson.ca/lib/ryerson/reader.action?docID=3433796&ppg=1.
Shanahan, Murray. “Whole Brain Emulation.” In The Technological Singularity, 15–50. Cambridge: The MIT Press, 2015b. https://ebookcentral-proquest-com.ezproxy.lib.ryerson.ca/lib/ryerson/reader.action?docID=3433796&ppg=75.
Shanahan, Murray. “Superintelligence.” In The Technological Singularity, 85-116. Cambridge: The MIT Press, 2015c. https://ebookcentral-proquest-com.ezproxy.lib.ryerson.ca/lib/ryerson/reader.action?docID=3433796&ppg=110.
Tamke, Martin, Paul Nicholas, and Mateusz Zwierzycki. “Machine Learning for Architectural Design: Practices and Infrastructure.” International Journal of Architectural Computing 16, no. 2 (2018): 123–43. https://doi.org/10.1177/1478077118778580.
Williams, Matt. “What Is A Singularity?.” Universe Today. January 7, 2017. https://www.universetoday.com/84147/singularity/.