Dennis Dollens Universitat Internacional de Catalunya, Barcelona. ESARQ [email protected]

AI-to-Microbe Architecture: Simulation, , Consciousness

Abstract Keywords

This paper probes questions of how big machines— Metabolic Architecture, Artificial Intelligence (AI), buildings—can function as hybrid metabolic/AI ALife, Alan Turing, Ludwig Wittgenstein, Computational organisms. Focusing on AI, artificial life (ALife), and Simulation. microbial intelligence I look through the lens of Ludwig Wittgenstein’s Tractatus and Alan Turing’s algorithmic 1. Propositions Toward Metabolic plant simulations to source modernist theory for Architectures biointelligent architectures. I’m using scant records and testimony interpreted through each thinker’s writings, Research strategies to regulate exploratory sets of architecture, and/or simulations. This text is then a generative design actions are called upon here for device for considering ways-of-being within, and ways- reasoning the inclusion of AI, synthetic life, and bio- of-thinking about, theory/practice for the fusion of algorithmic generation into the production of metabolic architectures. These exploratory tactics underwrite biological-to-biosynthetic (microbes, plants, animals, AI, machines.) Resulting theory thereafter hypothesizing biointelligent buildings as parts of nature. supports the development of bioremedial environmental Therefore, to link theory and observation I evolve cleanup addressing climate change. My proposition then strategies for the investigation of matter and forces deploys biomimetic and laboratory data to nurture starting with symbolic languages to sort types of metabolically driven intelligences partnered with AI intelligence. in the production of architectures. That ontological pathway stems from machine learning, bio-surveillance, Specifically, concepts-terms such as “atomic facts,” and digital simulation at object, agent, and urban scales. “form,” “objects,” “substance,” and especially “picture” Accomplishments in neural net AI and — autonomously inhabiting Wittgenstein’s Tractatus stirred me to question earlier breakthroughs in relation — are appropriated for research organization. His to current experimental practices. Subsequently, I link philosophy suits design analysis reinforced when the and hybridize emergent design proposition to AI, ALife, Vienna house [1] he designed for his sister is decrypted and biological intelligences as unities for environmentally to acknowledge Tractatus logic I analyzed in “Calculating performative, intelligent buildings. Turing, Thinking Wittgenstein” [2]. Likewise, for this paper, the Wittgenstein House stands as a pre-

70 Design and Semantics of Form and Movement Fig. 1. Brainstorming. Theory for plotting application and methodology for metabolic buildings involving microbe/plant/animal/ machine intelligences organized on Wittgenstein’s Tractatus and Turing’s Morphogen propositions. Drawn for the 2018 Metabolic Architectures Studio, UIC. Dennis Dollens. computer agent balancing thinking and design practice may then reveal biological attributes of living organisms to enact theoretical propositions in architecture. suited to architectural simulations and models. Viewed In the above framework, Wittgenstein’s philosophy through propositional tactics, the Wittgenstein House guides design-research, especially when joined by [1], designed approximately ten years after his book, notions of extended cognition [3], extended phenotypes is a built idiom — the logic of the Tractatus resolved [4], and Turing’s algorithmic botany [5]. That mix is in tectonic form: book-to-building, mind-to-matter. used to prompt theoretical organization to motivate Mind-to-matter scrutiny — mixing modes of autopoiesis AI-to-microbial [6] design conceptualization. Such [13] and extended cognition [3] — further situates the organization enables metabolic and AI linkages through house in a relationship with intelligence, language, and Wittgenstein’s logic and Turing’s simulations to illustrate visualization. relationships for theories of intelligent buildings. In the same way, designers may configure objectives for their 2. Typologies of Microbe, Plant, & Animal own projects, observations, and strategizing research Intelligence procedures for metabolic architectural [7. 8]. Metabolic architecture — formulated through propositions — articulate cognitive states supporting The result of Tractatus-prompted reasoning may be used to diagram design visualization theory enabling mind-to-matter design enactments. It seeks biological architects to consider: (i) the synthesis and mutability performance by questioning technological ways to of life, matter, and forces, (ii) differing typologies of extrapolate sensory biointelligence from nature. For intelligence existing in microbes, plants, animals, and example, asking: How can designers observe biological some machines for (iii) algorithmic simulation. All intelligence to enact architectural bioremediation? And, three areas have parallel interactions in Turing’s [9] if remedial strategies are sourced in nature: How can research that complement Tractatu [10] precepts when architects employ theoretical procedures to interpret used as foundational logic for AI-managed, metabolic an organisms’ intelligence and appropriate it to monitor architectures. Turing/Wittgenstein reformulations environmental toxicity? Such questions become thereby suggest pathways over which buildings may be recursive — referencing history and philosophy to designed as living technology [11. 12]. evolve thinking and data appropriate to building design. The designer may then set objectives receptive to living Together, Wittgenstein’s propositions and Turing’s metabolic operations involving the hybridization of AI, computational tactics enable design brainstorming [Fig. microbe, and synthetic biology. 1] to frame programs of investigation. That research

Design and Semantics of Form and Movement 71 In this phase, aggregates of AI/microbes may be studied or rendered in code to simulate microbial living as cellular-intelligent agents challenging architects to that, until now, have been genetically programmed only harness carbon dioxide (CO2) sequestration. First by nature [17]. Such observations and programming design explorations are organized through biology presuppose subsystems invoking collective microbe via autopoiesis (auto = self, poiesis = making) [13] behaviors that designers (working with biologists), interpreted to validate organisms (microbes, plants, must coax into an alliance with AI in order to ask: Can animals) as intelligent system unities. Theorized by buildings bio-technologically remediate pollution? In autopoiesis, organisms may be composite unities response, post-Turing questions take for granted that coupled with computational technology. Such intelligent bio-façades could incorporate, for example, simulations echo Turing’s plant observations for AI-monitored bacterial colonies (e.g., ) in order algorithmic performance [14] and his theories of to convert toxins to energy by feeding on CO2 in ways machine intelligence [15]. pioneered by oil-spill cleanup.

Shadowed by cybernetics, design research is linked to The thinking behind botanic algorithmic programming biology, biology is linked to code, and code is linked to and the exchange of metabolic data will enormously generative architecture. That architecture awaits new increase when AI and synthetic life are genetically questions formulated after Turing asked: “Can machines cooperating and reproducing in living matter found in think?” [9]. From Turing’s starting point, designers may nature [6. 26]. Existing examples preview beneficial fast-forward observations and data to contemplate living in animal architectures built of beeswax, intelligent systems programmed through computational wasp paper, biofilms, termite mounds, and human-made biology. In such cognitive-to-computational processes, adobe. Biocellular-AI may equally be modeled upon, or Tractatuslike corollaries emerge as tools [Figs. 1. 4] to paired with, microbes and plants to take residence in, determine design research methodologies situating and perform from buildings and urban infrastructures metabolic buildings as human-extended phenotypes [4. that track pollution while metabolically consuming 16]. specific toxins.

3. Hybridizing: Nature/Intelligence to 4. Turing/Wittgenstein Machine By extrapolating from Turing’s theories and Observation of nature’s biochemical processes reveals programming, I repurpose biology-to-code that cellular agents — microbes and plants — can be investigations through which he simulated aspects of integrated into synthetic materials or ALife provisioned matter, life, intelligence, and machine processing [5. 7]. for architectural components/facades. The resulting Procedurally, the lineage stems from his observations metabolic architectures operate as hosts for cellular, of living organisms extended to implant functions from living organisms communicating between urban nature into coding. This meshes with how Wittgenstein infrastructures and dynamic . For such hybrids, [10] used propositions to argue the “case” as the an architect needs biological data [Fig. 4] visualized and/ world and, in this text, when the Tractatus is culled for

Fig. 2. L-system Grown Plant/Microbe BioTowers. Left to Right: PagodaTower, BioTower, and MicrobeTower, Barcelona. Dennis Dollens.

72 Design and Semantics of Form and Movement Fig. 3. Left: feroxTowers are theoretical experiments whose function investigates hybrid AI/microbe atmospheric carbon capture collaboratively enabling options for bioremediation to produce energy at levels of cellular homeostasis. 2018-ongoing. Right: ArizonaTower STL & Animation Sequence. L-system plants generated as roots, branches, and seedpods — with the seedpods programmed as polysurface rectangles and the roots grown into a branching superstructure. Dennis Dollens.

design research logic. Extrapolated data, or mind’s- To unpack the above paragraphs requires us to: eye pictures in Wittgenstein’s sense, are then available (i) theorize technology, , architecture, and to researchers for cross-system investigations in computation in terms of bio/synthetic and metabolic which biochemical signals appear as potential codes propositions [Fig 1]. Doing so positions us to for heuristic AI to learn or decrypt. In my case-world conceptualize AI enhanced with animate intelligences experiments, algorithmic generation and morphological targeted for generative architecture while (ii) materialization [Figs. 2. 4] merge to help analyze tasking resultant theory to support programming teaching/research pathways over which metabolic bioarchitectural homeostasis engaged in climate, soil, architectures may source intelligences from nature for and water restoration. Those two processes integrate use in buildings [7. 8. 14]. networks of living (cellular organisms) and AI to (iii) envision metabolic intelligent buildings [Fig. 2]. 5. Biological Observation as a Design Operation The above bio-to-building dialectic [7. 8] enables us to characterize intelligent architectures as potentially Design research observations, channeled through 3D sentient and autonomous. One possibility is then scanning and SEM imagery [Fig. 4] thus parallel the to capture airborne CO2 through the actions of AI/ Tractatus’ first line: “The world is everything that is the metabolic machines performing intelligent analysis of case” [10]. Designers evolve individual case-worlds when toxins executed by microbes in biomechanical systems data is retrieved from organisms and applied to design [Fig. 3]. Sentience, routed to architectural functions, — inarguably a realm of human cognitive nature. In this algorithmic simulations, and cellular performance situation, scientific procedures supplement material and illustrates how metabolic machines could give rise environmental data to support design research. That to new species of design and architecture [Fig. 3]. research facilitates inducting ALife functionality and Emergent propositions thereafter expand territories of behaviors into building materials. Following preliminary experimental theory in queries such as: If metabolic/AI designs — data and imagery from microscopes and buildings sense, experience nature, and make decisions scanners [7. 8] — further detail the translation of through collective microbial life, do they experience insights [Figs. 1. 4] from nature in order to program AI/ artificial consciousness? [18. 19]. microbe material candidates into fablab productions and/or agents [Fig. 3].

Design and Semantics of Form and Movement 73 Fig. 4. Datura ferox Data/Image Sets. Top: Scanning Electron Micrographs of Datura ferox seedpod spikes. Middle, left: Datura ferox dried seedpod followed by two images of 3D-CT scans. Bottom: Rhino3D screen shots. Imported CT scans for carbon-capture pod development of (See: Fig. 3 Left). Dennis Dollens.

6. AI Anticipates DNA Turing Machines to themselves and, in a lesser register, to a few animal species [19. 24]. From such interrogations — not frequently probed, yet lurking behind, for example, Google’s AlphaGo, [20] — I These questions do not suggest a one-to-one microbe anticipate ontological nature-to-machine unity. Such or plantlike parity with deep learning or reinforcement questions are credible after Google’s AI succeeded at learning AI. Rather, I point out learning, playing, and winning Atari video games, beating [17] potential for genetic biotechnologies [12] to world champions at the Game of Go, and triumphing prompt computationally originated sensing in living/ over human and machine chess players [21. 22]. Still synthetic cells capable of next-generation inheritance other questions arise because programmers do not fully and reproduction. (According to Interface: The Royal understand the learning processes their codes engender Society — prospects include theoretically successful in machines. We/they may ask: Are some species of AI DNA Non-Deterministic Universal Turing Machines existentially thinking? [23]. Associatively: Is a subset [25. 26. 27]. Such molecular-scale machines would of neural net AI approaching cognitive abilities? — be compatible with quorum sensing and neural net abilities humans have traditionally considered exclusive

74 Design and Semantics of Form and Movement AI and could eventually be edited into living cells to Greek. The quandary — things seen cognitively — or enable DNA computing [26. 27]). Here then, metabolic cognitive things (res cogitans) — underpins design (ideas, buildings with active DNA intelligences, could be viewed propositions, prototypes) embodied in matter/tools as biosynthetic AI/ALife agents achieving microbe for thinking about metabolic architectures. Applied cognition long after Turing, but still consistent with to Turing, we need only look at his plant-to-algorithm his algorithmically simulated drawings and theoretical drawings and printouts [5] to comprehend that he writings [5]. seized on nature’s intelligence and physical growth for mathematic language and resulting digital simulations. To be clear, I am saying that Turing/Wittgenstein theoretical propositions may now be considered For design theory, this text’s case-world is in service members of intelligent nature. They are phenomenal to algorithmic simulation begun when Turing translated agents-of-thought licensed in philosophy, mathematics, plant attributes to programming [15. 28]. Simulation and cognitive science [3]. Propositions of this then enters our framework, not only through Turing phenomenal order (metabolic/living/AI intelligences) but also through Wittgenstein’s term to “picture” (res figure as agents of thinking [28] — facilitators from + idein). Similarly, propositions can be generative orders human cognition as it evolves new typologies of of simulation — Latin simulationem, simulare to underpin intelligence — in this text’s case-world — as metabolic thinking and designing as thought prior to coding. buildings and/or cities. Agents-of-thought, manifested through the Tractatus, are accordingly selected as design Turing simulated parts of nature in computation axioms for generating prototypes that consequently appropriate to philosophical design/machine/intelligence exist as extended phenotypes [4. 16]. debate. From that scenario, we confront results of the verb simulate and the noun simulation to communicate 7. Can Buildings Think? ranges of life/cognition. Simulations, for such usage, are thus human-extended phenotypes [4. 16]. They are However jarring, constructed species — neural net concept/objects of thought as cognitive or computed AI and ALife [11. 12. 27] — extend Turing’s question, numbers realized (built) as thinking machines/buildings “Can machines think?” [9. 14. 28] His question (and constructed in the world [3]. In autopoietic [13] terms, Wittgenstein’s too [29]), if answered positively, gives they are participants in cognitive-to-physical domains support to the proposition that buildings, as big (the case-world) that here includes Wittgenstein’s machines, can think. Design-research goals can then be theory of picturing [2. 10] incorporating language perused for ontological unity connecting ways-of-being and design/construction realized in symbolic logic / modes of debate / types-of-intelligence / responses- (philosophy) and the built Wittgenstein House [1. 2]. to-climate change / and requirements-of-design. As a result ontological cohesion aligns research with places, To give bearing to this paper, I see Turing as the tools, and nature as aspects of design contextualizing agent from whom we learned how to simulate nature the extension of our cognition [3] in case-worlds. with algorithms (e.g., his reaction/diffusion theory That process creates a framework for contemplating [5. 15]) as computational extensions of our thinking hybridized machines, AI, and (some) microbes as [3]. Those lessons later brought fourth code-to- environmental sentinels — new species of metabolic simulation languages (e.g., L-systems) for today’s output intelligence and artificial life [12]. whereby seeing (res) and imagining (ideate), translated mathematically, drive machines/AI to simulate nature. Metabolic architectures are thus first ideas, then With such ancestry in mind, I evolve models [Fig. 2] propositions (or codes) developed in design from using methods Turing pioneered. After extending his deductive exchange between phenomena, material, botanic observations for computational biology to computation, nature and the architect. In this lineage, CAD/CAM, I use laboratory and fabrication machines propositional analysis is realized descending, not only to visualize and build-out data resulting in various from the Tractatus, [10] but also from Latin res (thing) as scientific, technological, and design pathways for in res extensa; as well as idea found in idein (to see) from metabolic architectural practice [14].

Design and Semantics of Form and Movement 75 Consequently, by subscribing to precepts of Dawkins’s References [4] and Turner’s [16] extended phenotypes, I justify intermingling observations of organisms, matter, 1. Leitner, Bernhard. 1976. The Architecture of Ludwig and forces. Thereafter, metabolic architectures are Wittgenstein: A Documentation. New York. New York propositional descendants of Universal Turing Machines University Press. [25. 26] and/or botanic algorithms [5] theorized in The 2. Dollens, Dennis. 2019. Forthcoming” “Calculating Turing Chemical Basis of Morphogenesis [14] and Morphogen Thinking Wittgenstein: AI, ‘the Case,’ & Metabolic Theory of Phyllotaxis [27]. Architectures.” Virtual, Informal, and Built Landscapes. Bogota. Pontifica University. Prepublication: https://www.

Natural functions such as phyllotaxis in fir cones or academia.edu/37266358/Calculating_Turing_Thinking_ daisies, borrowed by Turing for his programming Wittgenstein_AI_the_Case_and_Metabolic_Architectures and drawings are now culturally and technologically 3. Clark, Andy. 2008. Supersizing the Mind: Embodiment, Action, absorbed. However, they may be further recast — and Cognitive Extension. New York. Oxford University computationally resimulated and biofabricated for Press. metabolic systems (living technology) not possible in 4. Dawkins, Richard. 1982. Extended Phenotypes: The Long his time [Fig. 2]. In that context, a paradox appears Reach of the Gene. New York. Oxford University Press. whereby we simulate nature in an act of intelligent/ 5. Dollens, Dennis. 2014. “Alan Turing’s Drawings, Autopoiesis phenomenal nature itself. Subsequently, at junctures of and Can Buildings Think.” Leonardo: The International observation other questions arise: If neural net AI can Society for the Arts, Sciences and Technology. simulate or partner with life, can it contemplate itself https://www.mitpressjournals.org/doi/abs/10.1162/ as conscious? [19]. Or: Can bioAI convince multicellular LEON_a_00766?journalCode=leon systems (microbes/plant cells) that its onboard 6. Falkowski, Paul G. 2015. Life’s Engines: How Microbes Made intelligence is a part of their living milieu? [12. 17. 18. Earth Habitable. Princeton. Princeton University Press. 19]. 7. Dollens, Dennis. 2018. “Studio: Digital Morphology for Metabolic Architectures.” https://uic-es.academia.edu/ 8. Conclusion DennisDollens 8. Dollens, Dennis, Estévez, Alberto T. 2017. "Metabolic Different iterations of neural net AI along with Architecture: Genetics, Digital Morphology, Turing, & AI". biological and synthetic life intelligences — are being In: VV.AA. Conference Proceedings, Kine[SiS]tem'17 - From developed and or decoded in laboratories around the Nature to Architectural Matter International Conference. world. Plant intelligence and communication predicated ISCTE - IUL, Lisbon (Portugal), June 19th-20th 2017. 248- on signaling and biochemical reaction/diffusion are 256. equally subjects of research [17. 31. 32]. We should 9. Turing, Alan M. 1951. “Can Digital Computers Think?” TS therefore interrogate composite bio/AI systems with AMT annotations of a talk broadcast on BBC Third [33] and dialectics [34], not through vague notions Programme. 15 May 1951. In: Cooper, S. Barry & van of sustainability, but through climate stabilization Leeuwen, Jan. Eds. (2013) Alan Turing: His Work and Impact. collectively underwritten by theories such as Amsterdam. Elsevier. 660. autopoiesis [13], extended cognition [3], and extended 10. Wittgenstein, Ludwig. 1922. Ogden, C.K. Tractatus Logico- phenotypes [4. 16]. Thereby, designers investigate Philosophicus. Kegan Paul. nature to pair with learning AI and/or living intelligence 11. Bedau, M. A. McCaskill, J. S. Packard, N. H. & Rasmussen, that address questions of how big machines — S. 2013. “Introduction to Recent Developments in metabolic architectures — can function like organisms Technology.” Artificial Life. The MIT Press. 19. 291-298. eradicating pollution while monitoring violations against 12. Zimmer, Carl. 2019. “Scientists Created Bacteria With a nature [35. 36]. Synthetic Genome. Is This Artificial Life?” 15 May 2019. https://www.nytimes.com/2019/05/15/science/synthetic- genome-bacteria.html?action=click&module=News&pgtype =Homepage 13. Maturana, Humberto & Varela, Francisco. 1980. Autopoiesis

76 Design and Semantics of Form and Movement and Cognition: The Realization of the Living. Dordrecht, January 2017. https://www.theguardian.com/science/2017/ Holland. D. Reidel Publishing Company. jan/23/organisms-created-with-synthetic-dna-pave-way-for- 14. Dollens, Dennis. 2017. Metabolic Architectures: Turing, new-entirely-new-life-forms Sullivan, Autopoiesis, & AI. Barcelona. Universitat 25. Turing, Alan M. 1936. “On Computable Numbers, With An Internacional de Catalunya. https://www.amazon.com/ Application To Entscheidungsproblem.” In: Cooper, S. Barry Metabolic-Architectures-Turing-Sullivan-Autopoiesis/ & van Leeuwen, Jan. Eds. (2013) Alan Turing: His Work and dp/1549954474/ref=sr_1_1?keywords=dollens+metabolic+a Impact. Amsterdam. Elsevier. 16. rchitectures&qid=1558989346&s=gateway&sr=8-1 26. King, Ross D. Currin, Andrew. Korovin, Konstantin. Ababi, 15. Turing, Alan M. 1952. “The Chemical Basis of Maria. Roper, Katherine. Kell, Douglas B. & Day, Philip J. Morphogenesis.” Philosophical Transactions of the Royal 2018. “Computing Exponentially Faster: Implementing a Society B. 237. 37-72. Reprinted in: Saunders, P.T. (1992) Non-Deterministic Universal Turing Machine Using DNA.” Collected Works of A.M. Turing: Morphogenesis. London. Interface: The Royal Society. London. North-Holland. 27. Dollens, Dennis. 2018. “Emergent Futures: Non- 16. Turner, Scott J. 2000. The Extended Organism: The Deterministic Universal Machines Using DNA — an Physiology of Animal-Built Structures. Cambridge, MA. Assessment for Metabolic Architectures.” AutopoietiX. Harvard University Press. https://autopoietix.blogspot.com/2018/11/emergent-futures- 17. Bessler, Bonnie, L. 2019. “How Bacteria Talk.” TedTalk. non-deterministic.html February 2009. https://www.ted.com/talks/bonnie_bassler_ 28. Turing, Alan, M. 1953. "Morphogen Theory of Phyllotaxis." on_how_bacteria_communicate In: Saunders, P.T. Ed. Collected Works of A.M. Turing: 18. Kuang, Ciff. 2017. “Can A.I. Be Taught to Explain Itself?” New Morphogenesis. London. North-Holland. York. The New York Times Sunday Magazine. 26 November 29. Wittgenstein, Ludwig. 1933/1965. The Blue and Brown 2017. Books. N.Y. Harper Perennial. 19. Dehaene, Stanislas. Lau, Hakwan. Kouider, Sid. 2018. “What 30. Langton, Christopher G. 1988/1989. “Artificial Life.” In: Is Consciousness, and Could Machines Have It?” Science. Langton, Christopher G. 1989. Artificial Life. Santa Fe, NM. 358. 486–492. 27 October 2017. http://science.sciencemag. Addison Wesley & The Santa Fe Institute. 6:1-47. org/ 31. Dollens, Dennis. 2015. “Autopoiesis + Extended Cognition 20. Burton-Hill, Clemency. 2016. “The Superhero of Artificial + Nature = Can Buildings Think?” Communicative and Intelligence: Can This Genius Keep It In Check?” London, UK. Integrative Biology. 31 August 2015. www.ncbi.nlm.nih.gov/ The Guardian. 16 February 2016. https://www.theguardian. pmc/articles/PMC4594259/ com/technology/2016/feb/16/demis-hassabis-artificial- 32. Pollan, Michael. 2013. “The Intelligent Plant: Scientists Debate intelligence-deepmind-alphago a New Way of Understanding Flora.” The New Yorker. 23 & 21. Sample, Ian. 2017. “It's Able To Create Knowledge 30 December 2013. 92-105. Itself': Google Unveils AI That Learns On Its Own.” The 33. Dollens, Dennis. 2019. “bioAI for Evolving Carbon Capture Guardian. 18 October 2017. https://www.theguardian.com/ Machines as Living Technology.” Preprint: https://uic-es. science/2017/oct/18/its-able-to-create- knowledge-itself- academia.edu/DennisDollens google-unveils-ai-learns-all-on-its-own 34. Engles, Frederick. 1964. Dialectics of Nature. London. 22. Sample, Ian. 2017. “Google’s DeepMind Makes AI Program Wellred Books. That Can Learn Like a Human.” The Guardian. 14 March 35. Stone, Christopher, D. 1972. 2000. Should Trees Have 2017. https://www.theguardian.com/global/2017/mar/14/ Standing: Law, Morality, and the Environment. Oxford googles-deepmind-makes-ai- program-that-can-learn-like- University Press. a-human 36. Chapron, G., Epstein, Y., & Lopez-Bao, J.V. 2019. “A Rights 23. Agence France-Presse, Shanghai. 2017. “World's Best Go Revolution for Nature.” Science. 29 March 2019. 363. https:// Player Flummoxed By Google’s ‘Godlike’ AlphaGo AI.” The science.sciencemag.org/content/363/6434/1392 Guardian. 23 May 2017. https://www.theguardian.com/ technology/2017/may/23/alphago-google-ai-beats-ke-jie- china-go 24. Sample, Ian. 2017. “Organisms Created With Synthetic DNA Pave Way For Entirely New Life Forms.” The Guardian. 24

Design and Semantics of Form and Movement 77