Exploring the connection between adaptive architecture and artificial intelligence

Interactive adaptive architecture refers to the design of systems that can adapt to changes in their operating environment through interaction with users or other agents in the system. Artificial intelligence, on the other hand, refers to the ability of computer systems to perform tasks that would otherwise require human intelligence. Interactive adaptive architecture and artificial intelligence are two interconnected concepts that have evolved with the development of cybernetics. The development of interactive architecture based on the use of artificial intelligence will not mean the disappearance of architecture, as we know it, but rather a change in the way it is perceived. In the future, architecture will be participatory, with an active role of inhabitants in shaping and transforming architectural habitats, but there will always be a need for a moderator, the architect, to design and set the rules of the game for humans.


Introduction
The continuous and accelerated state of change, reshaping and adaptation of the contemporary ethos creates an environment conducive to the emergence of forms of architecture based on the absorption of research and technological innovations from other fields.These constant and accelerating changes also directly imply the transformation of some aspects of how we think about the design and use of contemporary architectural objects.The speed of technological change is increasing; therefore, observation and analysis of the phenomenon are necessary in order to direct the effects, in a controlled way, to an area where they benefit man.The evolution of contemporary lifestyles therefore implies a rethinking of the paradigm by which the individual, architecture, technology and artificial intelligence relate to each other.
Today, buildings are designed to perform certain functions, but they are not always perfectly adapted to the needs of the users.They are designed to function in a certain way, which cannot easily be changed once the construction work is completed.These inconveniences therefore equate to excessive consumption of energy and resources, as well as reduced potential user comfort.The possibility of defining an architecture responsible for mediating interactions (dialogue) between humans and computers, i.e. automating certain processes necessary for reconfiguring the interior space and the way in which it relates to the exterior space, are relevant arguments for deepening the relationship between artificial intelligence and the built environment.Adaptive architecture aims to address these issues by creating buildings that can adapt to changes from the environment and to the needs of users, offering them a higher level of comfort and satisfaction.Adaptive architecture denotes a form of architecture capable of permanent self-determination in the face of changes from the environment.

Evolution history of the concept
The design of architectural space is constantly changing.In the past, this approach focused mainly on aesthetic, functional and ergonomic aspects, on meeting functional needs and adhering strictly to standards for creating quality architectural spaces, designed and built with care and skill.However, buildings were conceived as static structures, without taking into account the possible changes that IOP Publishing doi:10.1088/1757-899X/1304/1/012024 2 may arise from their use, in contrast to adaptive architecture which focuses on the implementation of flexible principles of design and use that allow the building to adapt to changes induced by the user or the environment.This new approach is based on the use of technology and intelligent systems to update the functional layout and interior climate according to the needs and requirements of the user.The history of this trend dates back to experimental projects in the 1960s and 1970s that explored the possibility of creating flexible, adaptable and reprogrammable spaces according to the needs and requirements of users.This new direction is based on the emergence of a literature of anticipation (science fiction), such as the writings [1] of J G Ballard, or on the development of systematic and critical research on the problems related to possible future developments of the society-technology binomial (futurology).
Indications of the imminence and magnitude of the transition to ubiquitous technologies were first presented by the American futurologist and writer A Toffler, best known for his book Future Shock (1970).He published a trilogy in which he warns of the negative psychological implications of accelerating technological change, described as "the future shock" or "the disease of change" [2].
In his essay "The Question of Technology", M Heidegger undertakes a search for the essence of technology, how it has affected man.M Heidegger believes that the instrumental definition of technology does not accurately present the truth and therefore cannot fully explain the complexity of the phenomenon.In his view, modern technology becomes an exploitation of nature, and thus a pure transformation of nature into a stock of raw materials subject to human orders, in contrast to traditional craftsmanship, in which the essence of technology [3] is closely linked to the rhythm of nature and the principles of art.The philosophical approach to contemporary technology provides an analysis of how it influences our perception of the world, and by implication how it shapes our context.
But the scientific community's concerns about the relationship between man and technology go back much further.In an attempt to find new ways of examining the complexity of intelligent behaviour in humans, animals and machines, the philosopher and natural historian G Canguilhem illustrated in his 1947 lecture "Machine et Organisme" [4] (first published in 1952) a different view of the living world, hitherto explained only on the basis of mechanistic philosophy and biology.It presents a new biological understanding of technology and machines as extensions of the human organism, explaining their construction in terms different from the simple application of scientific theorems or processes of industrialization and typification.Thus, G Canguilhem proposes an understanding of machines by referring to the structure and functions of organisms, which, unlike the quantifiable, predictable, calculable and periodic movements of mechanical systems, are characterised by self-regulation, self-maintenance, reproduction or polyvalence.This new vision is embodied in an emerging discipline of that period in the USA -bionics.The ideas presented above anticipated the research agenda proposed by the new science of cybernetics, founded by N Wiener in the 1940s.Just one year after (1948) G Canguilhem, Wiener presented the idea of informational feedback and selfregulation of systems as the basis for understanding biological, mechanical and social processes, suggesting the possibility of creating a new type of machine [5] capable of responding to the environment similar to living organisms.
The possibility of building machines more complex than the human body was considered a pipe dream until the mid-20th century.The 1940s and 1950s saw a series of attempts by scientists to find new ways of exploring the complexity of intelligent behaviour in humans, animals and machines by looking at message exchange -feedback-based communication -in both biological and technological systems.Mathematicians, physiologists and engineers such as J von Neumann, N Wiener, C Shannon, W McCulloch and M Mead synthesised information theory, communication theory and machine control theory in a series of conferences (Macy [6]) that ran from 1946 to 1953.They established the principles of cybernetics, marking the beginning of an intellectual revolution supported by the emergence of new fields of research such as artificial intelligence [7] (1956), cognitive science [8] and information and communication networks [9].The aim of cyberneticists was not to bridge the gap between organic and inorganic or natural and artificial scientifically; they did not propose a prototype simulating a living organism, but a new type of machine, capable at a theoretical level of moving beyond the opposition between the two poles.The adaptability of this new type of machine is given by the implementation of the "feedback loops" needed to ensure the predictability of the system's behaviour.Their role is to control the entropy of the system through feedback -the future behaviour of the system will be governed by its previous states.Thus, it has the ability to learn from previous experiences: at a theoretical level, this type of machine is equipped with sensors that record both external conditions and operating history, feeding this information back as input to the next selfregulating loop.
Self-regulation characterised systems in the first cybernetic wave (1945)(1946)(1947)(1948)(1949)(1950)(1951)(1952)(1953)(1954)(1955)(1956)(1957)(1958)(1959)(1960).In the biological world, this term was recognised as a property of homeostasis -the ability of living organisms to maintain an equilibrium by reducing energy losses in relation to changes in the environment.Researchers therefore believed that the homeostatic (first-order) cybernetic system was capable of reducing the loss of information from the system, while enhancing its control and stability through 'negative feedback'.Thus, a homeostatic machine will resist the tendency to disorder induced by informational noise in the environment (disturbances), maintaining the stability of its operating state by using feedback loops; in principle, their purpose is to correct all possible destabilising activities.Attempts to infuse architectural space with elements of cybernetics have resulted in defining a new potential quality of architectural space -adaptability.The phenomenon of architectural adaptability can materialise in many ways (through the use of removable elements, mechanical modules such as folding and sliding systems), but its character is not necessarily interactive, but rather static.From an architectural perspective, the first references to the phenomenon of adaptation through cybernetics or digital technologies can be found in the works of A Rabeneck, G Pask and C Eastman [10].
In 1969, A Rabeneck [11] proposed the incorporation of cybernetic technologies in order to create buildings that were as adaptable as possible to the future needs of their occupants; the expected result was to support both architects and future occupants by extending the life of a building.That same year, cybernetics pioneer G Pask outlined in his article "The Architectural Relevance of Cybernetics" [12] how cybernetics could induce evolutionary behaviour that responds and adapts to users' intentions and goals.Building on the earlier ideas of G Pask and A Rabeneck, C Eastman outlines the concept of "Adaptive-Conditional Architecture" [13], namely buildings capable of adapting according to feedback extracted from the environment (homeostatic model -first-order cybernetic system).It suggests that adaptability could be conceived as a self-adjusting feedback-based system, capable of dynamically managing spatial adjustments suited to the user's needs, similar to the way a boat constantly corrects its stability disturbed by changes in the environment (waves) [13].In 1975, N Negroponte exemplified this principle by means of an experimental greenhouse [14] equipped with glazing operated by thermostat-type mechanisms (of the cognitive node type), capable of controlling, decoding and processing information about the comfort level of plants.The operating scheme of a cognitive node is based on storing (in a database) and processing the identified requirements/needs of individuals.This database is accessed by a learning engine and a decision engine (made by repeated comparisons with already existing data).The recombination of information from the two leads to actions.The cycle repeats itself once the system receives feedback from the environment/user and rewrites certain information in the database.

Building interaction
The Extrapolating the phrase "to live" we deduce that a living architecture is made up of animated buildings that actively react to external stimuli, are capable of reproduction (but most importantly, self-sustaining).A 'living' architecture is one that is actively involved in larger ecosystems in which it interacts with other architectures, but also with other 'living' natural or artificial entities.Imagining such an architecture may at first evoke dystopian visions.
Popular examples of science fiction culture (such as the books/movies "2001: A Space Odyssey" [15], "I, Robot" [16], "Ex Machina" [17] or "Blade Runner" [18] ) bring into the focus of consciousness architectural spaces controlled by artificial autonomous technological entities.In countless such scenarios, protagonists are immersed in environments saturated with interacting "smart" devices (from flying machines and robots to ubiquitous interactive displays, sensors and control systems) -a future that, in contemporary society dominated by smartphones, drones, cloudcomputing, satellite constellations [19] and highly sophisticated military technology, does not seem too far away.Narratives presented in science fiction culture often involve the loss of human control over an artificial environment.
However, the influx of digital technology into architecture must be directed to the benefit of individuals and society as a whole.Cities 'populated' with 'smart buildings' that communicate with each other, able to modify and change their usage patterns, can generate constant feedback of information; thus 'smart and responsive' streets, walkways and public spaces could, through light, heat or cold, create microclimatic zones that attract and encourage certain human activities.The metabolizations behind these changes create different densities of activities that aim to redirect movement patterns.This 'intimacy' between the built environment, occupation of space and human movement can extend into the emergence of relationships between the urban/architectural landscape and society, energy and material resources.M Fox and M Kemp argue in Interactive Architecture [20] that interaction is a circular system -"'interacting', not just 'reacting'".
A clear definition of interactivity is fundamental.Even though the term 'interactive' is often overused in contemporary culture to define almost any IT product or service, it retains some ambiguity in common language.Interactivity is normally understood as the chain of interdependent bi-or multidirectional actions of two or more subjects.
The degree of interaction between subjects is directly proportional to the amount of information exchanged; inherently, such a relationship between interacting subjects is reciprocal, and they are altered in one way or another during the course of the interaction.Underlining the distinction that must exist between interactive and two-way or merely reactive communication, S Rafaeli proposes an interpretation of the notion of interactivity; according to him, interactivity is "the expression of the extent to which in a given series, any third (or subsequent) transmission (message) is related to the degree to which previous transmissions have referred" [21].
This seemingly complex description can easily be reduced to a classification of all types of communication that can occur between two or more subjects: • when messages are sent in two ways, but are not linked, bi-directional communication occurs; • if each of the messages sent only refers to a previously received message, we are dealing with reactive communication; • if this response is more complex and each message sent refers to several messages received in the past by the respondent, we are dealing with an interactive communication; thus, interactivity can be defined as a measurable, positive variable (>0) only when a communicator is able to relate his response to the information received before the one that triggers his present response.The value of interactivity increases with the number of (previously received) messages used to formulate responses.It is also important to stress the distinction between interactivity and interaction.If interaction is any action that takes place when two or more subjects interact, interactivity can be seen as the build-up of multiple interactions.Therefore, the notion of interactive (adaptive) architecture can be provisionally defined by buildings and constructed spaces capable of supporting an active dialogue with users and their environment -as T Jaskiewicz notes, the word "dialogue" is used metaphorically [22], as a reference to G Pask's Conversational Theory [23].
In the case of adaptive-interactive architecture, a dynamic action of a component could be seen not only as a reaction, but also as part of the overall interactivity scenario, therefore the distinction between interaction and reaction (i.e. the response of a system) is not clearly defined.Interactive architecture is a compelling alternative to simple building automation, with the potential to radically change the way buildings operate, are used and maintained.Applying the definition given by S Rafaeli, we deduce that automated spaces function purely reactively.Certainly, the automation of architecture has been an intermediate step in achieving adaptability through individual-built environment interaction, but at least two weaknesses can be identified in the automation of contemporary buildings from an interactivity perspective: • Automation systems are rarely integrated into the architecture of buildings, despite the more than obvious benefits they bring (economic, ecological, aesthetic etc.).[24] • Embedded systems have a linear logic, being built on chains of conditional statements ("if X, then Y, otherwise Z").Although building automation is by definition flexible, it cannot respond favourably to a wide range of dynamic factors as the system can only develop a response based on procedures anticipated by the system designers.They work optimally as long as the building is used exactly as it was designed to be used; in short, the ability to adapt to unforeseen conditions is poor or non-existent.However, the frequency of local adaptations to specific predictable factors is high.Replacing the linear logic guiding the behaviour of these systems with the ability to reason and learn autonomously (via Artificial Intelligence) can lead to interactivity.

(r)evolution of self-discovery
Artificial Intelligence (AI) as a discipline has already penetrated countless fields.However, the integration of AI into architecture is still in its infancy, but the results are promising and represent a potential reshaping of the discipline.Architecture is undergoing a truly profound (r)evolution, and technology is certainly one of the factors that will have a strong impact on architecture [25].Building design is already in a process of slow transformation both through the exploitation of new building technologies (e.g.3D printing, laser scanning, the use of physico-chemical properties of natural materials -wood -or synthetic materials -shape memory polymers etc.) and through sustained efforts to develop appropriate software or the implications of artificial intelligence in the built environment.The phrase "artificial intelligence" was coined by J McCarthy in 1955; it was needed to define the study of machines that exhibit and simulate "cognitive functions" (learning) or "problem solving" and the manifestation of intelligent behaviors [26].J McCarthy even believes that any notion in the sphere of learning or intelligence can be described precisely enough for a machine to simulate it.In contemporary times, artificial intelligence is considered the ability of a system to correctly interpret external data and learn from it to achieve specific goals through flexible adaptation".
One of the common definitions of artificial intelligence embedded in the building fabric can be seen as the ability of architectural space to exhibit emergent autonomous behaviour that emerges through computational processing of large data sets -in this context, any architectural object that develops its own ability to know its context (through sensors) and to interact dynamically and actively with the world can be considered intelligent [20].
The applications of Artificial Intelligence (AI) and Machine Learning (ML) in architecture are based on principles from cognitive science , which is why it is sometimes also called "cognitive architecture"."Artificial intelligent architecture" / "cognitive architecture" is a system capable of performing a range of cognitive tasks, including problem solving, as well as learning about aspects of its own performance.Currently, the demand for artificial intelligence and its influence in the building and architecture sector is low compared to other industries.Therefore, logically, the implementation of AI in architecture will transfer proven knowledge from other fields such as telecommunications, finance and commerce [28].
Artificial intelligence can be used both to optimise the performance and efficiency of buildings and to provide new ways of interacting with users and the environment; however, artificial intelligence embedded in adaptive architecture will not replace classical architecture and its paradigm, but will create a new discipline that will evolve alongside classical architecture.This may lead to a more adaptive and intelligent architecture.Classical architecture, based on aesthetic and cultural principles deeply rooted in the history and tradition of architecture, will continue to evolve and adapt to new technologies and trends.
The ubiquity of interactive adaptive architectural systems based on artificial intelligence has a significant impact on the cultural structure of society and on the individual in terms of psychological impact.In terms of society as a whole, this architecture may lead to an increase in individualism and a decrease in direct interpersonal interactions.Artificial intelligence-driven technology can give people a lot of options and possibilities, but it can also give them less motivation and a reduced ability to connect with others authentically.In addition, this can lead to increased feelings of discomfort and stress among people who feel overwhelmed by the amount of information and options available, which can lead to anxiety and social isolation.
As far as the individual is concerned, interactive adaptive architecture based on artificial intelligence can lead to a number of psychological impacts.On the one hand, it can lead to increased self-esteem and sense of control over life.For example, using AI-based technology to manage and optimise individual health can lead to an increase in self-esteem and confidence in one's ability to control own health and well-being.On the other hand, AI-based interactive adaptive architecture can lead to decreased ability to make decisions and adapt to unexpected changes.Due to the fact that these technologies are designed to provide suggestions and recommendations, the individual may become dependent to them and lose the ability to make autonomous decisions.
Thus, complex interactive adaptive architecture based on artificial intelligence has a significant impact on the cultural structure of society and on the individual in terms of psychological impact.It is important to consider both the social and individual impact of using AI-based technology and to take measures to minimise its negative effects and maximise its positive effects.

Conclusions
Interactive adaptive architecture and artificial intelligence are two interconnected concepts that have evolved in parallel with the evolution of cybernetics.Interactive adaptive architecture refers to the design of systems that can adapt to changes in their operating environment through interaction with users or other agents in the system.This approach has been developed mainly in the development of control systems and has been influenced by ideas from cybernetics, an interdisciplinary field concerned with the study of control and communication systems in living organisms and machines.Artificial intelligence, on the other hand, refers to the ability of computer systems to perform tasks that would otherwise require human intelligence, such as image recognition or natural language processing.This approach has been influenced by ideas from cybernetics and information theory, as well as developments in psychology and neuroscience.
The evolution of cybernetics has also influenced the evolution of interactive adaptive architecture and artificial intelligence through the development of models and algorithms inspired by biology and the study of complex systems.Thus, models of artificial neural networks and machine learning algorithms have been developed that allow systems to learn and adapt to changes in their operating environment.Natural language processing and image recognition technologies have also been developed, allowing systems to interact with users and interpret information from their operating environment.
Adaptive architecture and artificial intelligence present both considerable limitations and opportunities.Implementing artificial intelligence in architecture involves generating systems capable of learning and adapting to new situations, similar to the way the human brain develops and adapts.However, the whole phenomenon has a number of potential weaknesses: • Computing power: adaptive architecture that bases its operation on the use of artificial intelligence requires considerable computing power to process large volumes of data and perform complex calculations; • Training data: to obtain accurate and efficient results, artificial intelligence requires large amounts of diverse data to be trained.In certain domains or for more specific tasks, it can be difficult to create a sufficiently large, high-quality database.
However, we can also point to a number of promising predictions about the future evolution of the connections between adaptive architecture and artificial intelligence: • Increasing performance: technological advances and increasing computing power imply increasing efficiency and performance of adaptive architecture.Systems with integrated AI will be able to better process and understand complex information sets; • Increased interpretability: the evolution of AI and adaptive architecture will lead to the development of methods and techniques to improve the interpretability of AI.This will allow better understanding and justification of decisions made by AI -essential in areas such as health, security or housing; • Integration of other technologies: adaptive architecture and artificial intelligence can be combined with other technologies such as blockchain, IoT or virtual/augmented reality.
The development of interactive architecture is not equivalent to the disappearance of architecture as we know it.To some extent, much of today's architecture is already dynamic, interactive and adaptive.However, the speed and efficiency of current transformations are low.The development of interactive adaptive architecture is therefore equivalent to increasing the tempo of transformation.The revolution is therefore not in classifying adaptive architecture as a new type of architecture, but in changing the perception of any architecture as complex and adaptive.At the same time, the above statement does not imply that rapid spatial adaptation will necessarily become ubiquitous.On the contrary, scenarios can be imagined in which rapid architectural transformation is not and will never be necessary.Nevertheless, the complex adaptive view of architecture retains its validity and relevance.Adaptive architecture does not equate to the dystopian image in which future cities are full of dynamically changing or moving buildings.On the contrary, this development can equally guide us towards the development of seemingly static but actively optimising urban settlements only when these changes are explicitly desired by the city's inhabitants.
In conclusion, interactive adaptive architecture and artificial intelligence have been influenced by developments in cybernetics and information theory, as well as the study of complex systems in biology and other fields.The main common feature of all interactive spaces is the active role of inhabitants in shaping and transforming architectural habitats.Beyond any doubt, the future of architecture is a participatory one, focused on the ability of people to shape and improve the spaces in which they live their lives; but we should not assume that these can be created by users alone -there will always be a need for a moderator to design and set the rules of the game for people, while ensuring the development of enabling scenarios.It is up to the architect to take on this role.

Figure 1 .
Figure 1.A feedback loop: improving the operating parameters of the adaptive architecture by adding new datasets.

Figure 2 .
Figure 2. Conceptual architecture of a cognitive node

Figure 3 .
Figure 3.The information circuit -based on S. Rafaeli's diagram in "Interactivity: From New Media to Communication.", Sage Annual Review of Communication Research, 1988, 110-34