1

1 Artificial Intelligence in Landscape Architecture: A Literature

2 Review
3

4 1 Abstract

5 The use of artificial intelligence (AI) is becoming more common in landscape architecture. New

6 methods and applications are proliferating yearly and are being touted as viable tools for research

7 and practice. While researchers have conducted assessments of the state of AI-driven research

8 and practice in allied disciplines, there is a knowledge gap for the same in landscape architecture.

9 This literature review begins to fill the gap by searching and evaluating studies specifically

10 focused on AI and disciplinary umbrella terms (landscape architecture, landscape planning, and

11 landscape design). It includes searches of academic databases and industry publications that

12 combine these umbrella terms with the main subfields of artificial intelligence as a discipline

13 (machine learning, knowledge-based systems, computer vision, robotics, natural language

14 processing, optimization). Initial searches returned over 600 articles, which were then filtered for

15 relevance, resulting in about one hundred articles that were reviewed in depth. The work

16 highlights trends in dissemination, synthesizes emergent AI-Landscape (AI-LA) themes, and

17 argues for unifying dissemination and compilation in research and practice so as not to lose

18 relevant AI-LA knowledge and be caught off guard in the built environment profession’s next

19 technological leap.

20 2 Keywords

21 landscape architecture, landscape design, landscape planning, machine learning, optimization,

22 computational design
 2

23 1 Introduction

24 Leaders in landscape architecture have declared the need to consolidate data and

25 expertise from disciplines such as engineering, land planning, agriculture, and ecological

26 sciences to give “artistic physical form to modern ideals of equity, sustainability, resilience, and

27 democracy” (ASLA Is Committed to Climate Action, n.d.; New Landscape Declaration, 2016).

28 Such an assertion is fitting since landscape architects see their profession as an intersection

29 among all others dealing with spatial issues (Kullmann, 2016). As designers of all types of

30 exterior spaces, landscape architects’ work involves near-constant coordination with experts in

31 allied fields. This is especially evident in the current state of practice, where projects are

32 increasingly scaling up in scope to meet open-ended, territorial scale challenges (Bryant, 2021;

33 Polk, 2015). Yet, for all the diverse ways designers engage across disciplines, most simply lack

34 the time, knowledge, or background to account for the sheer number of ‘design problem’

35 permutations arising from multifaceted issues such as climate change resilience, large-scale

36 ecological degradation, and social equity. To this end, there is an emerging discussion around the

37 potential of artificial intelligence (AI) to address such limitations. The discussion includes topics

38 like laying a historical groundwork for AI (Z. Zhang, 2020), current and potential AI

39 applications to landscape architecture (Cantrell et al., 2021), proposing machine learning primers

40 and ontologies (Alina et al., 2016; Fernberg et al., 2021; Tebyanian, 2020), gauging the potential

41 for AI in coastal adaptation (Z. Zhang & Bowes, 2019), and conceptualizing an autonomous

42 post-human ecological infrastructure (Cantrell, Martin, et al., 2017).

43 Still, AI-focused literature remains underdeveloped in landscape architecture, leaving

44 knowledge seekers to turn to adjacent disciplines where the research is less nascent. The majority

45 of current research in AI systems for landscape design or planning focuses on either conceptual
 3

46 exercises or somewhat singular tools for specific applications. Even if current explorations evoke

47 broad observations about AI in landscape, a lack of compilation presents key unanswered

48 questions:

49 1) What exactly do we mean when we say AI in the context of landscape architecture?

50 2) How has AI been used in landscape architecture research/practice, if at all? And

51 3) Where are our current knowledge gaps with regard to AI?

52 This literature review seeks to lay a foundation to begin answering these questions. In it,

53 we: 1) establish a scope of review for landscape architecture and its subfields, 2) identify a

54 framework for artificial intelligence as a research area within which to embed the landscape

55 disciplines (i.e. the definition of AI as a discipline along with its sub-fields), 3) combine those

56 terms to perform a literature search using online databases, and 4) after refining results, we

57 provide a summary of trends, highlight emergent themes, and present the need for a future AI-

58 Landscape (AI-LA) research framework.

59 2 Defining Review Parameters

60 2.1 “Terms” of Landscape Architecture

61 Landscape architecture practice is interdisciplinary, so it can often be difficult to

62 delineate what falls under its purview. Grading, for instance, is a design exercise that can

63 reasonably be claimed by both engineers and landscape architects but is often taught, talked

64 about, and executed quite differently by each discipline. The same holds for many activities

65 landscape architects carry out (e.g. stormwater management, construction documentation,

66 landscape history, etc.). We recognize defining the scope of practice within landscape

67 architecture is integral for a comprehensive and systematic review of AI’s pervasion into the

68 entire discipline—and that such an undertaking could be enhanced by using established

 4

69 frameworks such as the Landscape Architecture Body of Knowledge (LABOK) survey findings

70 (2004) or Langley et al.’s knowledge domains of landscape architecture (2018). However, the

71 combination of these multi-level conceptual frameworks with the scope of artificial intelligence

72 is extremely vast. There have indeed been efforts to frame the context of the AI-LA knowledge

73 base (Cantrell et al., 2021; Tebyanian, 2020; Z. Zhang, 2020), but these works did not intend to

74 comprehensively review and formalize an AI-LA framework. Thus, for this review, we first

75 needed to establish a simple but encompassing disciplinary scope as the foundation for this

76 framework. We chose to adopt Ogrin’s definition of landscape architecture as a discipline which

77 comprises design and planning as two distinct subfields of creative work (1994). Hence our

78 scope uses the three disciplinary terms from Ogrin: landscape architecture, landscape design, and

79 landscape planning. These are often used interchangeably, and though sometimes seen as distinct

80 in detailed discussions of practice, they can confidently be lumped into a representative set that

81 represents the same discipline for the purposes of this review (von Haaren et al., 2014).

82 2.2 Artificial Intelligence and Applicable Subfields

83 The Oxford English Dictionary defines the term artificial intelligence (or AI) as “the

84 theory and development of computer systems able to perform tasks that normally require human

85 intelligence, such as visual perception, speech recognition, decision-making, and translation

86 between languages.” In the context of the AEC industry, the is often used colloquially as a catch-

87 all for highly technical or computational approaches toward design and automation. The term

88 machine learning is also used in common speak, often interchangeably with AI, even though it

89 technically represents only a subset of the AI field. The scope of AI is extremely vast, which has

90 led to the derivation of several subfields or branches. Here we outline some of the more common

91 subfields seen in literature to provide a framework for how we conceptualize the contributions
 5

92 and application of AI within landscape architecture. The primary subfields we explore in this

93 paper include: 1) Machine Learning, 2) Knowledge-based Systems, 3) Computer Vision, 4)

94 Robotics, 5) Natural Language Processing and 6) Optimization (Abioye et al., 2021; Public

95 Health Agency Canada, 2020). We acknowledge there is a range of other proposed subfields

96 (Chiabai et al., 2018; Mata et al., 2018; Zhu & Yan, 2015), but for this review we chose these six

97 as they are the most applicable to landscape architecture.

98 Machine learning. Machine Learning is one branch of AI, but the techniques often underpin a

99 range of different subfields. The term itself may often be used as a synonym for artificial

100 intelligence, perhaps because it is not well understood by non-experts or the diversity of AI

101 subfields is not well understood (and ever changing). In simple terms, machine learning focuses

102 on using statistical methods and models that can redefine and refine themselves to “learn.”

103 Learning is done through supervised, unsupervised or reinforcement learning. Supervised

104 learning necessitates a system to observe data, conduct analyses, and output to improve its

105 understanding of the analyzed phenomenon (Bzdok et al., 2018; Kotsiantis et al., 2007).

106 Unsupervised also uses statistical techniques which are suited to discovering patterns without

107 outputs or interaction with another agent such as a human or other computer system (Hastie et

108 al., 2009; Tarca et al., 2007). Reinforcement learning includes techniques where the computer

109 agent is intended to explore a set of actions or situations and then learn or anticipate outcomes

110 from different choice options (Sutton, 1992); the system learns the relationship between

111 consequence and action (Chandak et al., 2019; Huang, 2021). A simple example of machine

112 learning is an online application that learns purchasing habits and begins to make

113 recommendations based on your own patterns and those of individuals like you.
 6

114 Knowledge-based Systems (KBS). Knowledge-based systems are focused on using existing

115 knowledge to enable computational decision making. This subfield aims to develop inferences

116 about knowledge and enable user interaction to support, supplement or engage complex systems

117 (Akerkar & Sajja, 2009). These systems may require constructed representations of knowledge

118 (e.g. that use an ontology) with a particular focus on the relationship of the meaning of elements

119 within the set of knowledge. A KBS is an agent that adapts or creates inferences (Bergmann et

120 al., 2005) based on existing knowledge. While these systems have existed for some time, they

121 are not as popular given newer development in AI (Abdullah et al., 2006).

122 Computer Vision. Computer vision may be one of the more popular known AI techniques

123 within landscape architecture because of the subfield’s pursuits of simulating human perception

124 of visual elements (Szeliski, 2010). There are a range of approaches used in this subfield, with

125 some of the more recent oriented toward machine learning approaches. Computer Vision focuses

126 on pattern recognition (Chen, 2015) and object extraction (Prince, 2012). A popular tool

127 landscape architects use is Google Lens, which can identify a whole host of plants using

128 computer vision techniques.

129 Robotics. Robotics is centered on the use of sensors, often coupled with machine learning (often

130 reinforcement) and computer vision, to automate tasks. Robotics can encompass technology such

131 as autonomous vehicles (Faisal et al., 2019) and lawnmowers (Wasif, 2011), as well as systems

132 to irrigate and weed agricultural lands (Talaviya et al., 2020). Robotics can serve to replace

133 human actions but can also offer new forms of collaboration (Vrontis et al., 2022).

134 Natural Language Processing (NLP). Natural language processing is another subfield that

135 focuses on learning language and then recreating it to generate meaningful responses or outputs.

136 NLP uses a range of techniques to form an understanding of language, including grammar and
 7

137 lexicon, learning and language processing (statistical techniques), constructs and representation

138 (meaning and action), and techniques to manipulate language and learn the appropriateness of

139 those manipulations (Chowdhary, 2020).

140 Optimization. Optimization is another subfield within AI, that may often be misrepresented

141 within landscape architecture. While designers often attempt to optimize a given space, or

142 develop parametric models to aid in design, AI approaches necessitate some kind of learning or

143 algorithm to support the optimization. An important lesson here is that AI approaches usually

144 require a specific delineation of the problem in some quantifiable means. The techniques often

145 associated with optimization in AI are usually associated with search algorithms (Mirjalili &

146 Dong, 2020), such as genetic algorithms (Chamberlain & Meitner, 2009; Li et al., 2013),

147 simulated annealing (Rutenbar, 1989).

148 Importantly for all the subfields identified, the quantitative expression of constraints,

149 goals, inputs and outputs (when applicable) must be well defined. Fernberg and Chamberlain

150 (Fernberg et al., 2021) state that nearly every application of AI requires creating ontologies,

151 methods, data mining or expert-based learning and developing statistical approaches to facilitate

152 reasoning and may be done explicitly or implicitly. While humans play a range of defining roles

153 in AI, the key is that the machine is the learning agent. Learning happens, typically, with

154 abundant data, a clear language, and a reliable set of rules to follow.

155 3 Methodology

156 This section lays out a protocol for implementing our systematic review. In it, we

157 describe the process for searching, screening, and selecting literature that is sufficiently relevant

158 to the research objectives. Landscape Architecture encompasses activities and processes from a

159 range of disciplines. Many LA-related fields already have extensive AI-related literature reviews,
 8

160 such as urban forestry (César de Lima Araújo et al., 2021), urban design and planning (Abusaada

161 & Elshater, 2021; L. Yang et al., 2022), transportation (Abduljabbar et al., 2019), land use

162 planning (Chaturvedi & de Vries, 2021), horticulture (B. Yang & Xu, 2021), construction

163 (Abioye et al., 2021) and a range of others. For this review, we narrowed articles to specific

164 disciplinary keywords of Landscape Architecture, Design and Planning.

165 To be included in our review, articles must exist within a searchable English-based

166 literature database. All years of publication were included, though the recency of AI in literature

167 is relatively new (post 2000s). The initial literature search utilized three databases: Scopus,

168 IEEE, and JSTOR. Each of these was chosen to provide expansive interdisciplinary coverage

169 across the arts, humanities, and sciences—all of which are integral in some way to the landscape

170 and AI fields. JSTOR and a digital humanities affiliate called Constellate were used to find

171 landscape architecture industry insights, as JSTOR currently houses every issue of the official

172 periodical for the American Society of Landscape Architects (ASLA)—currently operating with

173 the moniker Landscape Architecture Magazine or LAM—from its first publication in 1910 up

174 until 2015. The most recent issues of LAM, from 2016 to the present, were searched and

175 screened using keyword searches on the publication website, URL

176 https://landscapearchitecturemagazine.org/. Hence, SCOPUS was chosen as the main data

177 source, while the others were used for full article download and data validation.

178 3.1 Search Strategy

179 The search terms comprised two lists, one encompassing all relevant AI techniques and

180 methods (and spelling modifiers) and one representing what we deem to be core landscape

181 discipline terms, organized into two single-line text strings then combined with the Boolean

182 operator AND. These terms were adapted from previous literature reviews of AI (Abioye et al.,
 9

183 2021; Emaminejad & Akhavian, 2022; Tebyanian, 2020; Wu & Silva, 2010; Yigitcanlar et al.,

184 2020) with some additional terms we added in order to be more exhaustive. We did not limit

185 applications of AI. The combination is as follows:

186 Line 1 (AI Search Terms): “Robotics” OR “Computer vision”, OR “Machine learning” OR

187 “Expert System” OR “Knowledge-based Systems” OR “Optimisation” OR “Optimization” OR

188 “Natural Language Processing” OR “Artificial Intelligence” OR “K-Means Clustering” OR

189 “Hierarchical Clustering” OR “Fuzzy Clustering” OR “Model-based Clustering” OR “Linear

190 Discriminant Analysis” OR “Monte Carlo” OR “Deep Belief” OR “Deep Boltzmann” OR “Deep

191 Learning” OR “Convolutional Neural Network” OR “Stacked Autoencoders” OR “Recurrent

192 Neural Network” OR “Deep Neural Network” OR “Speech processing” OR “Evolutionary

193 computing” OR “Evolutionary Algorithms” OR “Swarm Intelligence” OR “Discrete

194 Optimisation” OR “Convex Optimisation” OR “Discrete Optimization” OR “Convex

195 Optimization” OR “Automated Planning” OR “Ontology” OR “Automated Scheduling”

196 AND

197 Line 2 (Disciplinary Search Terms): “Landscape Architect*” OR “Landscape Design*” OR

198 “Landscape Plan*”

199 Scopus initially returned 528 results and IEEE returned 67. The search query could not be

200 effectively executed in the JSTOR database due to character limitations and a catalog method

201 which returned too many irrelevant results. We attempted to custom code our search using URL

202 hacks, but the results were still highly problematic. To ensure due diligence and not leave a

203 resource entirely, we attempted a simple Boolean-limited search using “Landscape Architecture”

204 and “Artificial Intelligence”. The initial return was >6000 results, and a quick browse of the first
 10

205 several dozens of these results found the included articles to be completely irrelevant to the topic.

206 However, after doing an advanced search in which the publication title had to contain the word

207 “landscape”, we were able to narrow the results to a return of 56 articles, three of which

208 contained a directly relevant subject matter (Lindhult, 1988; McCarthy & Portner, 1980; von

209 Wodtke, 1988). While these articles are not included in the formal results of our systematic

210 search, they will be touched on in the Discussion section. Furthermore, to account for other

211 sources that may not have been included in the systematic search process, we investigated

212 Google Scholar, Google. On Google Scholar and Google (web search) we used the same two

213 Boolean-limited search terms as used with JSTOR. These did not result in any substantially

214 different outcomes. Where possible, we included articles in the discussion.

215 3.2 Data Collection

216 Metadata and bibliographic information on the initial search results were exportable from

217 all databases and done so in two ways. The first was to export the saved searches in .RIS format

218 to Zotero reference management software, where each article’s bibliographic information along

219 with links to full text were organized into database-specific folders. The second data collection

220 method was an export of the saved searches into .CSV files, one from each database. The data

221 were then cleaned and combined into a common attribution structure joined into a single .CSV

222 file, which served as the principal dataset for our review and analyses. A cleaned table of the

223 data is included in Supplemental Materials.

224 3.3 Study Selection Coding

225 While the initial search returned a somewhat digestible literature chunk, it also returned

226 many duplicates and articles which seemed irrelevant to the purposes of this review—either

227 because the work did not constitute a true investigation of AI, did not utilize AI methods, or did
 11

228 not reasonably fall into the scope of landscape architecture/design or landscape planning, despite

229 the use of the Boolean operators to narrow the search.

230 To decide whether a study met the inclusion criteria of the review, we created a Python

231 script to further refine our master database. The code iterated through each item, by combining

232 the title, abstract and keywords and then identifying the frequency of keywords used that

233 matched our search terms. We used the same disciplinary search terms (“landscape architecture”,

234 “landscape design” and “landscape planning”) and then separated each of the subfields of AI

235 with their specific terms (each term listed was in quotes and shortened words utilized * for

236 Boolean limiting):

237 • Machine Learning: machine learning, supervised learning, unsupervised learning,

238 reinforcement learning, deep learning, k-means clustering, hierarchical clustering, fuzzy

239 clustering, model-based clustering, linear discriminant analysis, monte carlo, deep belief,

240 deep boltzmann, deep learning, convolutional neural network, stacked autoencoders,

241 recurrent neural network, deep neural network;

242 • Knowledge-based Systems: knowledge-based system, expert system, intelligent agent,

243 case-based reasoning, linked system, ontology;

244 • Computer Vision: computer vision, scene reconstruction, motion analysis, image

245 restoration, recognition;

246 • Robotics: robotic, climbing, actuation, locomotion;

247 • Natural Language Processing: natural language processing, speech processing, text

248 mining, text analy;

 12

249 • Optimization: optimiz, optimis, discrete optimi, convex optimi, evolutionary comput,

250 evolutionary algorithm, genetic algorithm, differential evolution, particle swarm, swarm

251 intelligence.

252 The script then coded each literature with the number of instances each of the disciplinary

253 terms and subfield keywords indicated in the matched fields, as well as a general search for

254 “artificial intelligence.” We further refined our data by eliminating any instances where no

255 keywords were present. This process provided a validation of the database search, by offering

256 complete control over the included literature. Further, as the script processed each literature row,

257 it identified if a duplicate article was found using year + title, since a DOI was not always

258 present. Duplicates were denoted in a separate file, then the authors manually confirmed and

259 removed them (85 in total).

260 Once all literature was coded, we then manually coded all dissemination venues (journal,

261 proceeding, book, etc.) for: 1) alignment to the disciplinary search terms and 2) review rigor of

262 the dissemination venue. Alignment of the field consisted of journals that are predominately

263 associated with the discipline, including adjacent journals or proceedings. For instance, venues

264 primarily aimed toward computer science or engineering were considered a low alignment for

265 LA. Further, review rigor was evaluated based on the reputation of the journal, including impact

266 scores (factors, cite score, etc.) and the review process. Coded values included: 1 = high

267 alignment and review rigor, 2 = combination of low/high or mid for both, and 3 = low alignment

268 and review rigor. These dissemination values (1-3) were then referenced with each article. The

269 full list of all venues and the tier scoring is provided in Supplemental Materials.

270 The resulting master dataset now provided a means to filter literature using:
 13

271 • Appropriateness of the venue and review rigor;

272 • Alignment with one or more of the disciplinary terms;

273 • An AI-related keyword.

274 The results and trends provided are delineated from different filtering mechanisms used.

275 The bulk of our commentary and detailed review of articles were from those with a score of 1 for

276 appropriateness of venue and review rigor, which also matched at least one disciplinary and AI

277 search term. These are referred to as tier 1 articles. We reviewed each filtered result and coded

278 them further across two additional criteria: degree of contribution and relevancy to the landscape

279 search terms. For the degree of contribution, we coded one of the following:

280 • Mention: merely mentions a disciplinary and AI term

281 • Discourse: theoretical or commentary

282 • Application: applies AI technique or approach

283 • Creation: develops new technique or heuristic

284 For relevancy, we denoted if an article seemed central to activities or knowledge related

285 to the landscape architecture discipline. There were instances where we recoded an article that

286 may have had a landscape-oriented search term but was completely irrelevant to AI, or vice

287 versa. Broader trends metrics include articles with a score of 2-3 for appropriateness of venue

288 and review rigor. These articles were not reviewed in depth and are referenced as tier 2 for the

289 purposes of this literature review. Tier 2 does not necessarily mean the contribution is of less

290 value, particularly if the article aligns primarily with other fields.

291 Further, we noted that articles with terms aligned with optimization were often not AI-

292 related, instead using the term to describe other quantitative or qualitative techniques. When used
 14

293 quantitatively, optimization overwhelmingly referred to a linear or stochastic technique to

294 optimize a space or design, typically with a set of environmental variables. Additionally, some

295 optimization articles focus on parametric modeling with mentions of optimization, but again

296 were clearly focused on the optimization of the model or design element without a coupled AI-

297 approach. We anticipate that several articles in tier 2 may be aligned with optimization, but not

298 with AI. After completing our search, we filtered all disciplinary results where optimization was

299 indicated without any other AI keyword. We then read through all titles to identify potential

300 articles that likely used AI techniques but may have not stated this explicitly or used a term that

301 may have been missed by our search terms. Any article we suspected may have used AI-coupled

302 approaches were flagged (roughly a dozen). Unfortunately, there precisely delineating the degree

303 to which AI is embedded across all optimization articles is nearly impossible. This is because

304 every article would need to be read in-depth (some of which are unavailable in full text) and

305 others with substantial interpretation (many have inadequate documentation of methods).

306 4 Review Results and Trends for AI-LA Applications

307 4.1 Results of Literature Review

308 A total of 600 articles were identified that met both the landscape keyword requirement

309 and the AI keyword requirement. These were published across over 300 different venues ranging

310 from top-tier journals, conference proceedings, individual university publications and book

311 publishers. Of the venues, 70 were tier 1 (priority for review), 31 were tier 2, and 207 were tier 3

312 (with 90% of those receiving the lowest ranking due to applicability to discipline and review

313 rigor). Of the 600 articles that met the tier 1 filter, 31 were associated with keyword “Landscape

314 Architecture”, 29 with “Landscape Design” and 150 with “Landscape Planning”, with ten of

315 these overlapping more than two of these terms.

 15

316 Upon reviewing all publications with keywords, the authors identified roughly one

317 hundred articles that meaningfully apply to the discipline and AI simultaneously and represent

318 the greater themes in the literature. The vast majority of these were application-based, with a

319 handful of others oriented towards theoretical or speculative discourse and a very select few

320 denoting a new advancement or creation.

321 4.2 Trends

322 The recent popularity and growth in AI-related works has been substantial. Figure 1

323 illustrates the rates of publication for each of the three disciplinary keywords. The figure shows

324 publications from 2000 to the end of November 2022 for all literature that met both AI and

325 disciplinary terms, as well as those literature published in top tier venues. As the chart indicates,

326 publications with the term “landscape planning” emerged earlier and was consistently producing

327 more than the other terms. While this is true for top tier venues, the trend has shifted recently

328 with “landscape design” emerging with more publications when all non-disciplinary venues and

329 lower tier venues were considered. From top tier venues, “landscape architecture” and

330 “landscape design” seem to have a similar output frequency with the latter slightly higher.

331 Broadly, the data show continued growth in the topic, with an extremely fast rise in publications

332 when considered all venues.

 16

333

334 Figure 1: Publication Counts of All Matching Keywords that met both discipline and AI keywords (2000-2022). Lines
335 show the results across tier 1 ranked dissemination publications (darker lines) and All tiers (lighter colors). X-axis is count of
336 publications.

337 Across all three terms, there were 12 publications before 2000, with the first in 1978 that

338 used a multiple hierarchical clustering method to help create a database of natural resources for

339 assessment and planning (Frondorf et al., 1978). The articles during this time period were

340 focused on database development, computer vision techniques and impact assessment. Some

341 were methodological (primarily within computer science venues) and other were applications

342 (primarily environmental journals). After 2000, there was a gradual increase in published works,

343 with the majority of works being published in the five years. In general, publications have

344 continued to rise across the umbrella landscape terms, with a significant drop during 2014-2016.

345 The most rapid rise has come since 2016.

346 It should be noted that in our review, the terms landscape design and planning

347 incorporated very broad definitions, with landscape design incorporating projects of a range of

348 areas, while planning was typically oriented toward larger areas. It was also more apparent that
 17

349 both landscape design and landscape planning were terms used in other disciplines when they

350 wanted to mention how their development or application of AI might align with other

351 disciplines. We noted that landscape architecture was not used as frequently in mentions, even

352 though the discipline does conduct both design and planning across scales.

353 We also identified author country affiliation across all publications. In total, we found

354 791 counts of country affiliations (meaning numerous articles were partnerships with scholars of

355 more than one country). Twelve countries were identified as having more than 10 affiliations

356 across all tiers, those countries are shown in Figure 2. Over one-third of the world’s countries,

357 with representation from all continents, have published something related to our search terms (67

358 countries). A full list of all affiliations is included in the Supplementary Documentation. The

359 rapid rise of AI-related publications across all tiers seems to emerge broadly across the world

360 with Chinese scholars leading this effort. It is important to recognize the substantial diversity of

361 projects and venues where authors publish – and the proportion of tier 1 to all tiers differs

362 substantially by country. Of the top 20 countries affiliated two thirds have about half of the

363 publications in a tier 1 venue, with over half of all countries publishing at least fifty percent of

364 articles in tier 1 venues. The overall trend indicates a growing interest in AI globally, which may

365 represent a likely increase in funding related to this work, the expertise necessary to

366 operationalize AI within the disciplines and partnerships being formed across disciplines.
 18

367

368 Figure 2: Author Country Affiliations showing the difference of affiliations by tier ranking.

369 4.3 AI Subfield Prevalence

370 We conducted an analysis of the distribution of AI techniques within the discipline

371 (landscape architecture, design and planning). The analysis observed all 600 publications that

372 returned one or more matching disciplinary keywords and AI keywords (including “artificial

373 intelligence”). Since artificial intelligence is not a single technique, for the purposes of reporting

374 here, we eliminated any article that did not mention one of the subtypes of AI. There were 62

375 instances where only “artificial intelligence” was used as a keyword without any other subtypes

376 indicated as a keyword. Of the 538 articles remaining, there were 597 total keywords instances

377 where one of the AI keywords was used (indicating several articles with more than one AI

378 subtype keyword included). The distribution of the subfields is provided in Figure 3.
 19

379

380 Figure 3: AI Subfield Distribution Counts of All Matching Keywords (discipline and AI)

381 Figure 3 demonstrates the vast proportion of works involve machine learning and

382 optimization, a pattern which mirrors that of other AEC industry disciplines (Abduljabbar et al.,

383 2019; Abioye et al., 2021). We investigated our data further, counting not only whether an article

384 mentioned a subfield, but also the total frequency of mentions of keywords. It is difficult to make

385 inferences about the meaning of the frequency of word use, but there is a slight increase in the

386 use of optimization and machine learning relative to the other subfields. This is likely because

387 most recent AI advancements have been within the realm of machine learning or optimization,

388 though this is quickly changing as fields natural language processing, robotics, and computer

389 vision are making exponential progress (Malone et al., 2020).

390 As we acknowledged earlier in the Methods, the keyword search for optimization

391 overestimates the number of contributions to literature in artificial intelligence because

392 optimization is a term that can be used qualitatively and parametrically where automated

393 learning is not central to the process but could be replaced with a stochastic or recursive

394 algorithm without learning. Subsequently, without having access to full text for all articles, we

395 conducted a review of titles and keywords manually to identify instances where optimization was

396 clearly indicating an AI technique. We found less than 5% of the optimization articles fit this
 20

397 criterion, but even after reviewing articles we could access in full text, it was not always clear if

398 their methods were actually AI because of limited documentation. As such, we have visualized,

399 in Figure 4, the distribution of all non-optimization techniques to emphasize the role of three

400 primary techniques used in the field. Likewise, the distribution of these subtypes through the

401 years (starting in 2000) is provided in Figure 5. This distribution shows a trend in the subtypes

402 that are associated with publications, suggesting machine learning and computer vision

403 applications have grown almost tenfold, whereas the other subtypes are dropping in proportion.

404 This is likely due to the increasing availability of tools and training scholars are using, as well as

405 the a natural shift away from other techniques (Abdullah et al., 2006).

406

407 Figure 4: Subset of AI Subfield Distribution Counts of All Matching Keywords (discipline and AI)
 21

408

409 Figure 5: Temporal distribution and use of different subfields of AI from Figure 4 (only showing 2000-2022)

410 4.4 Salient Themes in AI-LA Research and Practice

411 A close reading of the literature reveals significant themes in AI-LA knowledge work. These

412 themes range from a fine-grained focus on optimizing aesthetics or design process to using self-

413 improving algorithms for large-scale ecological modeling and forecasting, to analyzing policy

414 efficacy and public sentiment of open spaces through natural language processing. They are as

415 follows.

416 Design generation and evaluation. AI-driven applications for landscape design are proliferating

417 rapidly as landscape practitioners are learning how to extrapolate the technology to improve

418 design process and products. The review illustrates this occurring across a range of scales, from

419 Zhang et al.’s computer vision driven classification method for design details of Suzhou-style

420 private gardens (2021) to Naderi and Raman’s decision trees for pedestrian landscape designs

421 (Naderi & Raman, 2005), to a slew of academics and professionals’ use of machine learning for

422 generating concepts at the urban scale (Koma et al., 2017; Raman et al., 2022; Slager & De

423 Vries, 2013). There is also an emerging trend of AI applications for design evaluation, ranging

424 from improving machine perception of greenery (Suppakittpaisarn et al., 2022) to the use of
 22

425 computer vision, machine learning, and optimization techniques for post-occupancy evaluation

426 of user experience and ecosystem services in public open spaces (Schlickman, 2020; Wael et al.,

427 2022; X. Wang, 2021; J. Yang et al., 2022). Outside of the results found in academic databases,

428 our web searches revealed an abundance of AI-powered design applications being introduced or

429 operated. Some are directly relevant to landscape design, such as Autodesk and Sidewalk Labs’

430 tools for urban landscape design (Harrouk, 2020; Hickman, 2020); while others are more general

431 but have potential use and impact for design. These include apps like NVIDIA Canvas, which

432 allows users to make rough, color-coded brush strokes and instantly iterate them into landscape

433 renderings of various styles (Tack, 2021) and AI-powered text-to-image generators like

434 Midjourney, DALL-E 2, or Stable Diffusion, which create conceptual renderings from user-

435 generated text strings (Brezar, 2022; Dreith, 2022; Monge, 2022).

436 Perhaps the most obvious pervasion of AI applications into landscape architecture and

437 design workflows will be through the already burgeoning computational design ecosystem. In

438 2017, Proving Ground introduced LunchboxML, one of the first published plugins for machine

439 learning in the Grasshopper/Rhino3D environment (Miller, 2017), and a slew of ML plugins

440 have proliferated since. The following year, Cantrell and Mekies assembled a group of leading

441 professionals and academics to conjecture the role of parametric and computational design in

442 landscape architecture in a series of essays (2018), some of which anticipated a prompt pervasion

443 of AI applications into design (Ervin, 2018). The review results combined with perusal of non-

444 academic sources suggest such anticipation to be accurate, and also suggest the need for a better

445 way of documenting the phenomenon.

446 Ecological modeling. Computational ecology has been prolific in the AI literature, and the

447 field’s methods have begun to creep into modeling applications tooled for landscape design and
 23

448 planning purposes. For instance, Zhang and Bowes trained ML models that outperformed typical

449 models in real-time predictions of groundwater table response to storm surge in Coastal Virginia

450 (2019), and in turn posited a more machine-driven landscape monitoring regime. Abdollahi et al.

451 (2022) devised a new optimization approach to modeling urban ecosystem service zones based

452 on landscape patterns. On the other side of the urban-rural transect, Benke et al. introduce a

453 sophisticated application of geovisual analytics (driven by agent-based modeling) to model the

454 movements of ruminants in the landscape using satellite tracking data. While possibly not central

455 to the discipline as of yet, the concept of using advanced modeling to predict patterns of grazing

456 animals over large landscapes could be useful to consider as part of a design process. This is

457 especially true for animals that may use intentionally-designed large areas. Taking the idea of

458 machine-driven management further, Goodwin et al. (2022) and van Strien and Grêt-Regamey

459 (2022) both introduce ML methods for classification of landscape typologies. Taken with the

460 other autonomous management methods, a provocative question arises of whether AI utilization

461 could foster a land management regime that is entirely automated from start to finish.

462 There are also significant AI developments in forest planning and management. Salient

463 examples from the review include techniques to optimize (here we cite AI-optimization) for

464 timber harvest (Eyvindson et al., 2018; W.-Y. Liu & Lin, 2015), land use modeling (Lin et al.,

465 2009), habitat-specific restoration (Westphal et al., 2007), measuring forest connectivity (Peng et

466 al., 2019; Shanthala Devi et al., 2013) and spatial design of forests (G. Liu et al., 2006); machine

467 learning applications for species distribution modeling (Alegria et al., 2021; Ngarega et al.,

468 2021); modeling and planning for effects of fire in the forest landscape (Miranda et al., 2020;

469 Stamou et al., 2016; Zema et al., 2020); and modeling complexities of varied forest landscapes

470 (Ask & Carlsson, 2000; Gärtner et al., 2008; Hummel & Cunningham, 2006). These works
 24

471 represent only a sample of what has been done in Forestry—the discipline has been prolifically

472 producing optimization methods in recent decades (Kaya et al., 2016) and AI has creeped

473 significantly into urban forestry (César de Lima Araújo et al., 2021)—but are representative of

474 the research authors deemed relevant to landscape planning or design, whether in titles or

475 keywords.

476 Predictive analytics. Simulation and forecasting are another obvious anecdote for trending

477 methods in landscape and spatial planning, and the review gives evidence for it. Subjects cover

478 anything from using gaming technology, agent-based modeling (ABM) and AI to simulate

479 potential pedestrian and social life in urban spaces (Almahmood & Skov-Petersen, 2020) to

480 forecasting climate and emissions scenarios at the landscape scale (Bergier et al., 2019; Ngarega

481 et al., 2021), optimization for estimating green infrastructure potential (Dong et al., 2022), and

482 landscape simulations for improving predictive forest management (Hummel & Cunningham,

483 2006; Kampichler & Sierdsema, 2018; Stamou et al., 2016). While predictive analytics only had

484 a handful of results falling under the umbrella term of “landscape planning”, the fact that they

485 are among the most common methods in AI-driven urban planning, internet of things (IoT) or

486 Smart Cities conceptualizations (Souza et al., 2019) makes them very relevant to the landscape

487 disciplines, as many decisions and models will inevitably creep into the operational territory of a

488 landscape architect or planner focused on urban environments.

489 Landscape policy evaluation. A number of studies utilized AI methods to model ecosystem

490 services. For instance, Groot et al. used evolutionary algorithms for generating planning and

491 design solutions for multi-functional landscapes (2018); Queiroz et al. used k-means clustering

492 to map and classify ecosystem services bundles (2015); while others modeled socio-ecological

493 determinants, associations, or natural capital stocks and flows associated with ecosystem services
 25

494 (Lorilla et al., 2020, p.; Mouchet et al., 2014; Zank et al., 2016). Other projects utilized AI as

495 part of evaluating landscape policy outcomes (both potential and actual). These include

496 Berkhardt et al., who used machine learning to generate land use classifications from remote

497 sensing imagery in order to measure conformity to and impacts of water conservation measures;

498 Wang et al.’s Monte Carlo simulation technique to measure cooling and energy saving potentials

499 of shade trees and urban lawns in Phoenix (2016); clustering methods for prioritization of green

500 corridor development (Shapira et al., 2013); and development of machine learning tools for

501 maximizing biodiversity benefits in conservation planning (Thomson et al., 2020).

502 Sentiment analysis and social media. Sentiment analysis (SA), or sentiment modeling, is a

503 burgeoning research area that uses text and image data mining and to understand public opinion

504 of issues, services, or social phenomena, among other things (L. Zhang & Liu, 2017). The

505 methodology has grown precipitously over the last decade and pervaded across a wide variety of

506 fields, mostly due to the abundance of user data generated in social media (Yue et al., 2019). The

507 landscape and urban design disciplines are included in this creep (C. Yang & Liu, 2022), and

508 review results suggest future growth as public engagement methods evolve among researchers

509 and practitioners. Much of the work to date centers around public green space satisfaction. Song

510 et al. utilized computer vision (including face and object detection models) to analyze and

511 annotate imagery captured from social media platforms to inventory and assess characteristics

512 such as temporal patterns of park use, social dynamics, activities, and demographics (2022).

513 Jahani et al. applied artificial intelligence techniques to identify the prevalence of bird sounds in

514 urban green spaces and their association with mental restoration (2021). Ghermandi et al.

515 extracted online geolocated photographs from social media platforms then used computer vision

516 cloud services to characterize human-open space interactions in urban green spaces (2022).
 26

517 Wang et al. zoomed out to a regional scale as they employed machine learning techniques to

518 assess green space satisfaction of 50 parks in Beijing (2021). They also introduced a landscape-

519 feature lexicon to help improve granularity of landscape sentiment analysis. Other studies focus

520 on measuring sense of place in important cultural or touristic landscapes such as the Las Vegas

521 Strip, USA (Song et al., 2021) or Mt. Huangshan, China (Chai et al., 2021), or on simply

522 understanding discrepancies between policy measures and user experience using natural

523 language processing of user-generated text data (Wartmann et al., 2021).

524 Knowledge systems for AI-LA applications. Another less prolific but important grouping of

525 studies are theoretical or speculative pieces touching on the permeation of AI methods into

526 landscape practice and the need to formulate knowledge frameworks that help designers and

527 planners adapt to it. Zhang provides a historical sketch of cybernetic environments, positing that

528 landscape designers have previously had influence on their development and should reclaim that

529 influence to drive the future (2020). Cantrell et al. argue through synthesis of current

530 developments that AI’s fast-growing influence presents an epistemological crisis for landscape

531 architecture and that the profession may need to rethink its authorial role in solving wicked

532 problems of the day (2021). In accordance with this frame, Fernberg et al. suggest addressing the

533 crisis involves formalizing operational language into ontological frameworks for AI systems

534 (2021) and that there is a need to grow more systematic knowledge of AI in landscape

535 architecture. Exemplary efforts to do so include Tebyanian’s review and primer for machine

536 learning in urban landscape design (2020) and Ervin’s history and taxonomy of digital landscape

537 architecture, which gives historical context to computational developments and associated

538 progression in landscape architecture while providing commentary about terminology and
 27

539 definitions—including one of the first references in the literature to the concept of ‘bionic’

540 landscapes (2020).

541 5 Discussion

542 In carrying out the review process, the authors drew some distinct impressions of the

543 state of AI in landscape architecture. Broadly, sentiment toward AI within the field is growing

544 rapidly. This is depicted by the diversity of AI-based implementation across all publications, the

545 global distribution of work and likely the recognition of the importance of design from within

546 more computationally centric fields. Yet even amongst the most non-technical, discipline-

547 focused venues for landscape architecture, planning, and design, there appears to be an uptick in

548 publications. Further, the sophistication and implementation of AI methods may demonstrate the

549 increased training and access to techniques that are being afforded researchers, as well as

550 funding opportunities globally. Importantly, researchers within the discipline who are interested

551 in AI should become aware of the vast interest from other disciplines who want to engage in the

552 discipline, in particular being aware that much of the growth in the topic is associated with the

553 term “landscape design”. More broadly, we reflect on Fernberg and Chamberlain (2019) who ask

554 about the role technology specialists might play within the future of landscape architects. To

555 what extent will landscape architects (here we speak more broadly toward designers and planners

556 as well) develop and embrace AI taking agency on how it is implemented within the discipline,

557 or will technology designers from outside the discipline shape the discipline using AI?

558 It is important to underscore that the while scope of this review focuses on direct

559 relevance to the umbrella terms “landscape architecture”, “landscape design”, and “landscape

560 planning”, the breadth and depth of AI-related research increases significantly with the inclusion

561 of terms or activities that could feasibly fall under the umbrella of the landscape architecture
 28

562 discipline but have greater relevance or recognitions in allied fields or disciplines. For example,

563 research advancements of automation in agriculture and ecology are longstanding, and now

564 converging to offer unique solutions to global food security problems. Researchers have seen

565 success in applications ranging from vegetation biomass and cover estimation in fire-damaged

566 landscapes (Anderson et al., 2018), measuring forest tree defoliation using smart-phone photos

567 (Kälin et al., 2019), or using image-based deep learning models for disease detection in

568 agriculture (Mohanty et al., 2016) to thermal mapping waterbodies, forest monitoring, and aerial

569 seeding using UAS (Amorós & Ledesma, n.d.; Hogan et al., 2017; Minařík & Langhammer,

570 2016; Novikov & Ersson, 2019; Sai et al., 2020; Vovchenko et al., 2020). Combining artificial

571 intelligence (AI) applications in agriculture with emergent methods in agroecology shows the

572 potential to address pressing problems in 21st century food systems such as climate change

573 uncertainty, optimizing data flows, or crop efficiency (Barbieri et al., 2018; Cherkauer et al.,

574 2018; Leippert et al., 2020). Most if not all of these applications have some relevance to

575 landscape architecture or landscape planning—as some designers work in agricultural contexts

576 or are interested in applications for ecological restoration in their site planning—but the subjects

577 of the studies in and of themselves may not be considered central to the practices, teachings, or

578 research of landscape architecture.

579 Another interesting area of convergence that may appear less obvious is in robotics.

580 While the literature search only returned one article on robotics in the landscape disciplines—

581 Westort and Shen’s exploration of robot-assisted, in-situ landscape gardening (2017)—the

582 authors see robotics as an emerging theme. The exponential growth of robotics in the AEC

583 industry as suggested by Abioye et al. (2021) and Emaminejad and Akhavian (2022), the man

584 established architectural robotics labs (International Map of Robots in the Creative Industry,
 29

585 n.d.), and an uptick in landscape-oriented robotics projects from institutions such as Louisiana

586 State University and ETH Zurich (Harmon et al., 2022; Hurkxkens et al., 2020, 2022; Johns et

587 al., 2020)—projects not picked up in the literature search because of term mismatch—there is

588 clear evidence that this subfield of AI has potential for an outsized impact on the landscape

589 disciplines, particularly design.

590 While a distinction between relevant AI research in agriculture or robotics and landscape

591 design is fairly intuitive, the line becomes thinner when considering fields like urban design and

592 urban planning, which overlap significantly with landscape disciplines in interests, theory, and

593 methods (Van Assche et al., 2013). For instance, there are a number of extensive and already

594 highly cited reviews of artificial intelligence in urban planning subjects such as land planning

595 dynamics (Wu & Silva, 2010), planning for smart cities and big data (Allam & Dhunny, 2019;

596 Yigitcanlar et al., 2020), transportation planning (Abduljabbar et al., 2019), and urban forestry

597 (César de Lima Araújo et al., 2021; Nitoslawski et al., 2019). All of these have direct relevance

598 to landscape design in urban contexts but would be otherwise unknown in a review that only

599 includes the keywords “landscape architecture,” “landscape design,” or “landscape plan”—

600 which could in turn mean hundreds of informative studies on landscape-relevant AI applications

601 go unnoticed from parochial scoping in terms.

602 Furthermore, the same dilemma applies to the more specialized terms of landscape

603 architecture. If, for example, a reader would rely on the current study which focuses more

604 broadly on the discipline, they would consider AI development to be overwhelmingly nascent

605 with just a few dozen relevant studies. But if they were to perform a search using “stormwater

606 management,” one of the specializations of which licensed landscape architects are required to

607 have some knowledge, they would find an abundance of well-established literature on AI
 30

608 applications for stormwater plans (Imran et al., 2013). In the authors’ view, this exercise paints a

609 complicated picture wherein the vast majority of contributions to AI development relevant to

610 landscape architecture come from researchers and practitioners outside the discipline; a paradox

611 where AI-LA research and practice is at once established and emerging, quite possibly to the

612 ignorance of many in the profession in either sense. Such a notion suggests that practice-based

613 researchers should be aware that using only discipline-specific terminology in precedent research

614 could unintentionally blind them to relevant information if they are too parochial in keyword

615 usage. On the other hand, a more robust output of AI-LA research from within the discipline

616 could bolster the relevance of its lexicon and help to avoid constant borrowing and fitting of

617 knowledge from outside it. In other words, the knowledge domain unique to landscape

618 architecture could effectively build a new appendage that relates to AI and its use in practice and

619 scholarship.

620 Given these limitations, we suggest that future work can more comprehensively

621 illuminate the role of AI in landscape research and practice by expanding the scope of the

622 research and utilizing a broader but systematic lexicon of disciplinary terms. For example, a

623 future study could include a full-scale systematic literature review that takes the current work’s

624 AI search terms protocol and queries literature using established disciplinary frameworks such as

625 the Landscape Architecture Body of Knowledge (LABOK Task Force, 2004) or the core

626 landscape knowledge domains developed by Langley et al. (2018). Doing so could likely provide

627 a more encompassing panorama of AI-related work that includes the facets of the profession that

628 clearly fall under its purview but do not always carry the labels of “landscape architecture”,

629 “design”, or “planning”. Besides expanding the terminology, future AI-LA reviews or other

630 investigations should also seek to bridge the knowledge accessibility gap between academia and
 31

631 practice. While the current work illustrates practice-driven AI research and applications as

632 published in the industry standard Landscape Architecture Magazine and white papers from a

633 handful of practice-based research labs, the question of how to appropriately (and systematically)

634 compile knowledge from industry and synthesize it with academic literature remains largely

635 unsolved. A protocol for addressing this problem will provide mechanisms for consistent and

636 defensible longitudinal research on AI’s transformations of the profession in coming decades.

637 As part of this special issue in Landscape Journal, we set out to explore how artificial

638 intelligence has and is influencing landscape architecture, design, and planning. In conducting

639 this review one of the more difficult decisions was selecting the bounds of a discipline, that is, by

640 definition, rather interdisciplinary. Those reading this article are likely to have read and most

641 certainly will read articles from a variety of different disciplines that relate or conduct research

642 on landscapes. In many contexts the definitions of architecture, design and planning within

643 landscape often blend, especially when referenced from outside the discipline. Ironically, in our

644 search we not only discovered the increase in AI-related publications within these fields of study

645 and practice, but a significant body of literature published in venues and by authors outside of

646 these disciplines that give mention to their potential contribution to one or more of these three

647 landscape terms. However, the wide range of different publication venues cataloged from our

648 search and ranking techniques makes it difficult to ascertain the role AI might play within the

649 discipline in the future. This is because most of the articles associated with the discipline come

650 from lower tier venues where stated relevance to practice and research are vague.

651 The question of what defines landscape architecture, or landscape design or landscape

652 planning is an ontological and socio-cultural question. In our section, “Terms” of Landscape

653 Architecture we provide some context for why we set out to identify these three terms and to
 32

654 ascertain the contribution of AI within these narrower definitions of what these fields practice.

655 We discovered an increasing trend of AI-related publications in venues central to these

656 disciplines and that the rapid rise of this work has surged in the past few years. From within

657 landscape architecture the rise has only increased recently. For instance, in the 2022 issue of the

658 Journal of Digital Landscape Architecture, the authors identified several new publications that

659 applied artificial intelligence techniques, with some of those being direct applications and others

660 referencing the significance of the techniques (Barbarash et al., 2022; Fengjing et al., 2022;

661 Khalilnezhad, 2022; X. Liu & Tian, 2022; J. Yang et al., 2022).

662 One of the significant challenges of this research endeavor was identifying if and to what

663 extent AI is playing a role in practice and education. Most literature reviews, including our own,

664 often focus on peer-reviewed publications, or at a minimum, dissemination products that show

665 up in literature related databases. Unfortunately, outside of Landscape Architecture Magazine

666 (LAM) and the LAF Case Studies repository, there are not any obvious centralized venues for

667 publishing practice-oriented work, at least in the US. While LAM has published AI-related

668 articles (Cantrell, Ellis, et al., 2017; Fernberg & Chamberlain, 2021; Petrich, 1986; Zeiger,

669 2019), these are limited in number and primarily contributions from academic scholars. We ask

670 whether or not this is an indication of the lack of AI-related work being conducted in practice or

671 if there is a knowledge and dissemination gap. As discussed in emerging themes, we are aware of

672 several efforts from landscape architecture practice involving AI applications, but these

673 contributions are not being included in searchable databases. Such a lack of compilation can

674 make identifying contributions from practice very difficult and limit the democratization of these

675 works, even if that is not the intent.

 33

676 It is at the intersection of disciplinary recognition, ontology and the dissemination of

677 works from the fields identified that we see a conundrum. Does landscape architecture, design

678 and planning play a key role in proliferating or at least applying AI-related work? Are scholars

679 within the field publishing in other disciplinary journals and not giving credit to the contribution

680 to their field or is dissemination not taking place, or is there really a limited amount of work? In

681 any case, we argue that researchers and practitioners should consider including search terms that

682 relate to the broader landscape disciplines, while also including AI-related keywords in abstracts

683 and metadata associated with publications. This may help to raise awareness of the contributions

684 within the field and bring greater recognition to the application of these techniques to other

685 disciplines, as well as make this information more readily available to students, practice and

686 scholars. A specific example of this could be the use of the term “landscape design”.

687 Interestingly, it appears the overwhelming increase in publications across all venues is associated

688 with this term but come from venues outside the discipline. Further, in the articles we reviewed

689 that used this term, we noticed that it often serves as a catch-all that might be more appropriately

690 delineated as landscape architecture or landscape planning. Thus, in an effort to promote our own

691 disciplinary contribution toward AI, future publications may want to consider adding ”landscape

692 design” to keyword searchers where publications are AI centric. This may increase the likelihood

693 of knowledge sharing within and outside landscape centric disciplines. When considering the

694 general pulse of publications across all venues, the relative growth and access of AI-related

695 techniques shows plausible continued growth of AI-related articles.

696 6 Conclusion

697 After reviewing hundreds of articles, websites, books, and proceedings, we believe our

698 observations can be reasonably summed up in three important takeaways:

 34

699 1. Interest and contributions toward AI are growing steadily and significantly in the

700 landscape discipline, both in academic research and professional applications.

701 2. Applications and discourse from all subfields of AI have grown exponentially over the

702 past three years. This, in our view, suggests the emergence of a new technological

703 paradigm for the discipline.

704 3. Landscape researchers in all sectors (e.g. academia, practice, government) would be well

705 served to formalize, compile, and contribute to a clear AI-LA knowledge framework

706 and/or AI-LA standards of practice to ensure proper workforce preparedness (whether in

707 pedagogical or professional settings).

708 4. To promote AI knowledge sharing across all disciplines, more universally accepted terms

709 (e.g. landscape design), should be included in AI publications within the discipline.

710 5. The need for scholars and practitioners to improve the democratization of knowledge

711 sharing by ensuring publications are indexed and easily accessible (e.g. open-source)

712 from a variety of databases (e.g. Google Scholar, Scopus).

713 Engagement with technology driven by artificial intelligence, both practically,

714 speculatively, and critically, is increasing year over year in landscape architecture, design, and

715 planning, and will continue to do so. This literature review is the first attempt at providing a

716 formal epistemic baseline for said engagement and incite a more systematic approach to

717 compiling the knowledge it produces. As artificial intelligence systems continue to permeate

718 everyday landscape practice, the workforce will have to confront a number of adaptive

719 challenges. How and where do we integrate AI into existing design and planning processes? Do

720 those processes fundamentally change because of said integration? How will landscape
 35

721 practitioners ensure that the AI systems mediating their workflows are producing socially and

722 environmentally equitable outcomes? We argue that such questions can only be answered if there

723 is a formal framework for understanding how AI has, does, and will affect the state of practice.

724 The review shows evidence that AI-LA knowledge is nascent even if rapidly growing, hence

725 current gaps in the literature could be reasonably identified or filled with a more systematic

726 method for measuring AI’s influence in the more detailed subsets of landscape disciplines,

727 especially one that bridges dissemination gaps between academia and professional practice. If

728 researchers, professionals, and educators act now to develop this protocol, it could serve as

729 leverage for landscape to take the lead in shaping a techno-vernacular of the future. If we

730 hesitate, we run the risk of causing unnecessary root shock to the profession because of failure to

731 get ahead of the next technological tipping point AI is pushing us towards.

732 7 References

733 1. Abdollahi, S., Ildoromi, A., Salmanmahini, A., & Fakheran, S. (2022). Optimization of

734 geographical space of ecosystem service areas and land-use planning, Iran. Environmental

735 Monitoring and Assessment, 194(8), 527. https://doi.org/10.1007/s10661-022-10204-7

736 2. Abduljabbar, R., Dia, H., Liyanage, S., & Bagloee, S. A. (2019). Applications of Artificial

737 Intelligence in Transport: An Overview. Sustainability, 11(1), Article 1.

738 https://doi.org/10.3390/su11010189

739 3. Abdullah, M. S., Kimble, C., Benest, I., & Paige, R. (2006). Knowledge-based systems: A

740 re-evaluation. Journal of Knowledge Management.

741 4. Abioye, S. O., Oyedele, L. O., Akanbi, L., Ajayi, A., Davila Delgado, J. M., Bilal, M.,

742 Akinade, O. O., & Ahmed, A. (2021). Artificial intelligence in the construction industry: A
 36

743 review of present status, opportunities and future challenges. Journal of Building

744 Engineering, 44, 103299. https://doi.org/10.1016/j.jobe.2021.103299

745 5. Abusaada, H., & Elshater, A. (2021). Competitiveness, distinctiveness and singularity in

746 urban design: A systematic review and framework for smart cities. Sustainable Cities and

747 Society, 68, 102782. https://doi.org/10.1016/j.scs.2021.102782

748 6. Akerkar, R., & Sajja, P. (2009). Knowledge-Based Systems. Jones & Bartlett Publishers.

749 7. Alegria, C., Roque, N., Albuquerque, T., Fernandez, P., & Ribeiro, M. M. (2021).

750 Modelling maritime pine (Pinus pinaster aiton) spatial distribution and productivity in

751 Portugal: Tools for forest management. Forests, 12(3). Scopus.

752 https://doi.org/10.3390/f12030368

753 8. Alina, P., Oliviu-Dorin, M., Iuliana, B., & Valean, H. (2016). Developing a Feasible and

754 Maintainable Ontology for Automatic Landscape Design. International Journal of

755 Advanced Computer Science and Applications, 7(3).

756 https://doi.org/10.14569/IJACSA.2016.070351

757 9. Allam, Z., & Dhunny, Z. A. (2019). On big data, artificial intelligence and smart cities.

758 Cities, 89, 80–91. https://doi.org/10.1016/j.cities.2019.01.032

759 10. Almahmood, M., & Skov-Petersen, H. (2020). Public space public life 2.0: Agent-based

760 pedestrian simulation as a dynamic visualisation of social life in urban spaces. Journal of

761 Digital Landscape Architecture, 2020(5), 305–317. Scopus.

762 https://doi.org/10.14627/537690032

763 11. Amorós, L., & Ledesma, J. (n.d.). Aerial Robotics for Forest Management and Seeding. 10.

764 12. Anderson, K. E., Glenn, N. F., Spaete, L. P., Shinneman, D. J., Pilliod, D. S., Arkle, R. S.,

765 McIlroy, S. K., & Derryberry, D. R. (2018). Estimating vegetation biomass and cover
 37

766 across large plots in shrub and grass dominated drylands using terrestrial lidar and machine

767 learning. Ecological Indicators, 84, 793–802. https://doi.org/10.1016/j.ecolind.2017.09.034

768 13. Ask, P., & Carlsson, M. (2000). Nature conservation and timber production in areas with

769 fragmented ownership patterns. Forest Policy and Economics, 1(3–4), 209–223. Scopus.

770 https://doi.org/10.1016/s1389-9341(00)00016-2

771 14. ASLA Is Committed to Climate Action. (n.d.). Asla.Org. Retrieved December 18, 2020,

772 from https://www.asla.org/climateaction.aspx

773 15. Barbarash, D., Rasheed, M., & Gupta, A. (2022). Automated Recording of Human

774 Movement Using an Artificial Intelligence Identification and Mapping System. Journal of

775 Digital Landscape Architecture. https://doi.org/10.14627/537724007

776 16. Barbieri, L., Wyngaard, J., Galford, G. L., Thomer, A., Bittner, C., & Adair, C. (2018).

777 Technology in Agroecosystems: Integrative Land-Atmosphere Monitoring with Unmanned

778 Aerial Systems. 2018, IN13B-08.

779 17. Bergier, I., Silva, A. P. S., Abreu, U. G. P. D., Oliveira, L. O. F. D., Tomazi, M., Dias, F.

780 R. T., Urbanetz, C., Nogueira, É., & Borges-Silva, J. C. (2019). Could bovine livestock

781 intensification in Pantanal be neutral regarding enteric methane emissions? Science of the

782 Total Environment, 655, 463–472. Scopus. https://doi.org/10.1016/j.scitotenv.2018.11.178

783 18. Bergmann, R., Kolodner, J., & Plaza, E. (2005). Representation in case-based reasoning.

784 The Knowledge Engineering Review, 20(3), 209–213.

785 https://doi.org/10.1017/S0269888906000555

786 19. Brezar, Z. (2022, October 26). Using Artificial Intelligence In Your Design Process

787 [Online Magazine]. Landezine. https://landezine.com/using-artificial-intelligence-in-your-

788 design-process/
 38

789 20. Bryant, M. (2021). Learning Spatial Design through Interdisciplinary Collaboration. Land,

790 10(7), 689. https://doi.org/10.3390/land10070689

791 21. Bzdok, D., Krzywinski, M., & Altman, N. (2018). Machine learning: Supervised methods.

792 Nature Methods, 15(1), 5.

793 22. Cantrell, B., Ellis, E., Hill, K., & Martin, L. (2017). Ecology On Autopilot. Landscape

794 Architecture Magazine, 107(6).

795 http://bt.royle.com/article/Ecology+On+Autopilot/2787563/409226/article.html

796 23. Cantrell, B., Martin, L. J., & Ellis, E. C. (2017). Designing Autonomy: Opportunities for

797 New Wildness in the Anthropocene. Trends in Ecology & Evolution, 32(3), 156–166.

798 https://doi.org/10.1016/j.tree.2016.12.004

799 24. Cantrell, B., & Mekies, A. (2018). Codify: Parametric and Computational Design in

800 Landscape Architecture. Routledge.

801 25. Cantrell, B., Zhang, Z., & Liu, X. (2021). Artificial intelligence and machine learning in

802 landscape architecture. In The Routledge Companion to Artificial Intelligence in

803 Architecture. Routledge.

804 26. César de Lima Araújo, H., Silva Martins, F., Tucunduva Philippi Cortese, T., & Locosselli,

805 G. M. (2021). Artificial intelligence in urban forestry—A systematic review. Urban

806 Forestry & Urban Greening, 66, 127410. https://doi.org/10.1016/j.ufug.2021.127410

807 27. Chai, C., Song, Y., & Qin, Z. (2021). A thousand words express a common idea?

808 Understanding international tourists’ reviews of mt. Huangshan, china, through a deep

809 learning approach. Land, 10(6). Scopus. https://doi.org/10.3390/land10060549

 39

810 28. Chamberlain, B. C., & Meitner, M. J. (2009). Automating the visual resource management

811 and harvest design process. Landscape and Urban Planning, 90(1), 86–94.

812 https://doi.org/10.1016/j.landurbplan.2008.10.015

813 29. Chandak, Y., Theocharous, G., Kostas, J., Jordan, S., & Thomas, P. (2019). Learning

814 Action Representations for Reinforcement Learning. Proceedings of the 36th International

815 Conference on Machine Learning, 941–950.

816 https://proceedings.mlr.press/v97/chandak19a.html

817 30. Chaturvedi, V., & de Vries, W. T. (2021). Machine Learning Algorithms for Urban Land

818 Use Planning: A Review. Urban Science, 5(3), Article 3.

819 https://doi.org/10.3390/urbansci5030068

820 31. Chen, C. H. (2015). Handbook Of Pattern Recognition And Computer Vision (5th Edition).

821 World Scientific.

822 32. Cherkauer, K. A., Bowling, L. C., Lee, C., Singh, N., Smith, S., & Zhu, Y. (2018). A multi-

823 scale approach to improved simulation of agroecosystem response to climate in the

824 Midwestern United States. 2018, H34H-09.

825 33. Chiabai, A., Quiroga, S., Martinez-Juarez, P., Higgins, S., & Taylor, T. (2018). The nexus

826 between climate change, ecosystem services and human health: Towards a conceptual

827 framework. Science of The Total Environment, 635, 1191–1204.

828 https://doi.org/10.1016/j.scitotenv.2018.03.323

829 34. Chowdhary, K. R. (2020). Natural Language Processing. In K. R. Chowdhary (Ed.),

830 Fundamentals of Artificial Intelligence (pp. 603–649). Springer India.

831 https://doi.org/10.1007/978-81-322-3972-7_19
 40

832 35. Dong, J., Guo, F., Lin, M., Zhang, H., & Zhu, P. (2022). Optimization of green

833 infrastructure networks based on potential green roof integration in a high-density urban

834 area—A case study of Beijing, China. Science of The Total Environment.

835 https://doi.org/10.1016/j.scitotenv.2022.155307

836 36. Dreith, B. (2022, November 16). How AI software will change architecture and design

837 [Online Magazine]. Dezeen. https://www.dezeen.com/2022/11/16/ai-design-architecture-

838 product/

839 37. Emaminejad, N., & Akhavian, R. (2022). Trustworthy AI and robotics: Implications for the

840 AEC industry. Automation in Construction, 139, 104298.

841 https://doi.org/10.1016/j.autcon.2022.104298

842 38. Ervin, S. M. (2018). Turing landscapes. In B. Cantrell & A. Mekies (Eds.), Codify (1st ed.,

843 pp. 89–114). Routledge. https://doi.org/10.4324/9781315647791-9

844 39. Ervin, S. M. (2020). A brief history and tentative taxonomy of digital landscape

845 architecture. Journal of Digital Landscape Architecture, 2020(5), 2–11. Scopus.

846 https://doi.org/10.14627/537690001

847 40. Eyvindson, K., Rasinmäki, J., & Kangas, A. (2018). Evaluating a hierarchical approach to

848 landscape-level harvest scheduling. Canadian Journal of Forest Research, 48(2), 208–215.

849 Scopus. https://doi.org/10.1139/cjfr-2017-0298

850 41. Faisal, A., Kamruzzaman, M., Yigitcanlar, T., & Currie, G. (2019). Understanding

851 autonomous vehicles: A systematic literature review on capability, impact, planning and

852 policy. Journal of Transport and Land Use, 12(1), 45–72.

 41

853 42. Fengjing, L., Dong, L., & Liwei, X. (2022). Assessing the Green View Index in Chinese

854 Cities: An Example with Data from Eighty Cities. Journal of Digital Landscape

855 Architecture. https://doi.org/10.14627/537724028

856 43. Fernberg, P., & Chamberlain, B. (2021, August 19). I, Designer? Landscape Architecture

857 Magazine, 111(8). https://landscapearchitecturemagazine.org/2021/08/19/i-designer/

858 44. Fernberg, P., Sturla, P., & Chamberlain, B. (2021). Pursuing an AI Ontology for Landscape

859 Architecture. Journal of Digital Landscape Architecture, 2021(6), 452–460. Scopus.

860 https://doi.org/10.14627/537705040

861 45. Frondorf, A. F., McCarthy, M. M., & Rasmussen, W. O. (1978). Data-intensive spatial

862 sampling and multiple hierarchical clustering: Methodological approaches toward cost/time

863 efficiency in natural resource assessment. Landscape Planning, 5(1), 1–25. Scopus.

864 https://doi.org/10.1016/0304-3924(78)90013-8

865 46. Gärtner, S., Reynolds, K. M., Hessburg, P. F., Hummel, S., & Twery, M. (2008). Decision

866 support for evaluating landscape departure and prioritizing forest management activities in

867 a changing environment. Forest Ecology and Management, 256(10), 1666–1676. Scopus.

868 https://doi.org/10.1016/j.foreco.2008.05.053

869 47. Ghermandi, A., Depietri, Y., & Sinclair, M. (2022). In the AI of the beholder: A

870 comparative analysis of computer vision-assisted characterizations of human-nature

871 interactions in urban green spaces. Landscape and Urban Planning, 217. Scopus.

872 https://doi.org/10.1016/j.landurbplan.2021.104261

873 48. Goodwin, C. E. D., Bütikofer, L., Hatfield, J. H., Evans, P. M., Bullock, J. M., Storkey, J.,

874 Mead, A., Richter, G. M., Henrys, P. A., Pywell, R. F., & Redhead, J. W. (2022). Multi-tier
 42

875 archetypes to characterise British landscapes, farmland and farming practices.

876 Environmental Research Letters, 17(9), 095002. https://doi.org/10.1088/1748-9326/ac810e

877 49. Groot, J. C. J., Yalew, S. G., & Rossing, W. A. H. (2018). Exploring ecosystem services

878 trade-offs in agricultural landscapes with a multi-objective programming approach.

879 Landscape and Urban Planning, 172, 29–36. Scopus.

880 https://doi.org/10.1016/j.landurbplan.2017.12.008

881 50. Harmon, B. A., Nam, H. Y., Gilbert, H., & Iravani, N. (2022). Living Typography:

882 Robotically Printing a Living Typeface. CHI Conference on Human Factors in Computing

883 Systems Extended Abstracts, 1–4. https://doi.org/10.1145/3491101.3519894

884 51. Harrouk, C. (2020, December 8). Spacemaker Proposes AI-Powered Generative Design to

885 Create More Sustainable Spaces and Cities. ArchDaily.

886 https://www.archdaily.com/952850/spacemaker-proposes-ai-powered-generative-design-

887 to-create-more-sustainable-spaces-and-cities

888 52. Hastie, T., Tibshirani, R., & Friedman, J. (2009). Unsupervised Learning. In T. Hastie, R.

889 Tibshirani, & J. Friedman (Eds.), The Elements of Statistical Learning: Data Mining,

890 Inference, and Prediction (pp. 485–585). Springer. https://doi.org/10.1007/978-0-387-

891 84858-7_14

892 53. Hickman, M. (2020, October 15). Sidewalk Labs launches Delve, a generative design tool

893 for optimized urban development. The Architect’s Newspaper.

894 https://www.archpaper.com/2020/10/sidewalk-labs-launches-delvegenerative-design-tool-

895 for-optimized-urban-development/
 43

896 54. Hogan, S. D., Kelly, M., Stark, B., & Chen, Y. (2017). Unmanned aerial systems for

897 agriculture and natural resources. California Agriculture, 71(1), 5–14.

898 https://doi.org/10.3733/ca.2017a0002

899 55. Huang, L. (2021). Design of landscape ecological environment monitoring system based on

900 improved particle swarm optimization. Fresenius Environmental Bulletin, 30(6), 6207–

901 6214. Scopus.

902 56. Hummel, S., & Cunningham, P. (2006). Estimating variation in a landscape simulation of

903 forest structure. Forest Ecology and Management, 228(1–3), 135–144. Scopus.

904 https://doi.org/10.1016/j.foreco.2006.02.034

905 57. Hurkxkens, I., Fahmi, F., & Mirjan, A. (2022). Robotic Landscapes: Designing the

906 Unfinished. Park Books.

907 58. Hurkxkens, I., Mirjan, A., Gramazio, F., Kohler, M., & Girot, C. (2020). Robotic

908 Landscapes: Designing Formation Processes for Large Scale Autonomous Earth Moving.

909 In C. Gengnagel, O. Baverel, J. Burry, M. Ramsgaard Thomsen, & S. Weinzierl (Eds.),

910 Impact: Design With All Senses (pp. 69–81). Springer International Publishing.

911 https://doi.org/10.1007/978-3-030-29829-6_6

912 59. Imran, H. M., Akib, S., & Karim, M. R. (2013). Permeable pavement and stormwater

913 management systems: A review. Environmental Technology, 34(18), 2649–2656.

914 https://doi.org/10.1080/09593330.2013.782573

915 60. International Map of Robots in the Creative Industry. (n.d.). Association for Robots in

916 Architecture. Retrieved July 15, 2022, from https://www.robotsinarchitecture.org/map-of-

917 creative-robots
 44

918 61. Jahani, A., Kalantary, S., & Alitavoli, A. (2021). An application of artificial intelligence

919 techniques in prediction of birds soundscape impact on tourists’ mental restoration in

920 natural urban areas. Urban Forestry and Urban Greening, 61. Scopus.

921 https://doi.org/10.1016/j.ufug.2021.127088

922 62. Johns, R. L., Wermelinger, M., Mascaro, R., Jud, D., Gramazio, F., Kohler, M., Chli, M., &

923 Hutter, M. (2020). Autonomous dry stone. Construction Robotics, 4(3), 127–140.

924 https://doi.org/10.1007/s41693-020-00037-6

925 63. Kälin, U., Lang, N., Hug, C., Gessler, A., & Wegner, J. D. (2019). Defoliation estimation

926 of forest trees from ground-level images. Remote Sensing of Environment, 223, 143–153.

927 https://doi.org/10.1016/j.rse.2018.12.021

928 64. Kampichler, C., & Sierdsema, H. (2018). On the usefulness of prediction intervals for local

929 species distribution model forecasts. Ecological Informatics, 47, 67–72. Scopus.

930 https://doi.org/10.1016/j.ecoinf.2017.07.003

931 65. Kaya, A., Bettinger, P., Boston, K., Akbulut, R., Ucar, Z., Siry, J., Merry, K., &

932 Cieszewski, C. (2016). Optimisation in Forest Management. Current Forestry Reports,

933 2(1), 1–17. https://doi.org/10.1007/s40725-016-0027-y

934 66. Khalilnezhad, S. T. (2022). Using Twitter as a Means of Understanding the Impact of

935 Distance and Park Size on Park Visiting Behavior (Case Study: London). Journal of Digital

936 Landscape Architecture. https://doi.org/10.14627/537724015

937 67. Koma, S., Yamabe, Y., & Tani, A. (2017). Research on urban landscape design using the

938 interactive genetic algorithm and 3D images. Visualization in Engineering, 5(1). Scopus.

939 https://doi.org/10.1186/s40327-016-0039-5
 45

940 68. Kotsiantis, S. B., Zaharakis, I., & Pintelas, P. (2007). Supervised machine learning: A

941 review of classification techniques. Emerging Artificial Intelligence Applications in

942 Computer Engineering, 160(1), 3–24.

943 69. Kullmann, K. (2016). Disciplinary convergence: Landscape architecture and the spatial

944 design disciplines. Journal of Landscape Architecture, 11(1), 30–41.

945 https://doi.org/10.1080/18626033.2016.1144668

946 70. LABOK Task Force. (2004). Landscape Architecture Body of Knowledge Study Report.

947 American Society of Landscape Architects.

948 https://www.asla.org/uploadedfiles/cms/education/accreditation/labok_report_with_append

949 ices.pdf

950 71. Langley, W. N., Corry, R. C., & Brown, R. D. (2018). Core Knowledge Domains of

951 Landscape Architecture. Landscape Journal, 37(1), 9–21. https://doi.org/10.3368/lj.37.1.9

952 72. Leippert, F., Darmaun, M., Bernoux, M., & Mpheshea, M. (2020). The potential of

953 agroecology to build climate-resilient livelihoods and food systems. FAO and Biovision.

954 https://doi.org/10.4060/cb0438en

955 73. Li, X., Lin, J., Chen, Y., Liu, X., & Ai, B. (2013). Calibrating cellular automata based on

956 landscape metrics by using genetic algorithms. International Journal of Geographical

957 Information Science, 27(3), 594–613. https://doi.org/10.1080/13658816.2012.698391

958 74. Lin, Y.-P., Verburg, P. H., Chang, C.-R., Chen, H.-Y., & Chen, M.-H. (2009). Developing

959 and comparing optimal and empirical land-use models for the development of an urbanized

960 watershed forest in Taiwan. Landscape and Urban Planning, 92(3–4), 242–254. Scopus.

961 https://doi.org/10.1016/j.landurbplan.2009.05.003
 46

962 75. Lindhult, M. S. (1988). The Road Beyond CAD. Landscape Architecture, 78(5), 40–45.

963 JSTOR.

964 76. Liu, G., Han, S., Zhao, X., Nelson, J. D., Wang, H., & Wang, W. (2006). Optimisation

965 algorithms for spatially constrained forest planning. Ecological Modelling, 194(4), 421–

966 428. Scopus. https://doi.org/10.1016/j.ecolmodel.2005.10.028

967 77. Liu, W.-Y., & Lin, C.-C. (2015). Spatial forest resource planning using a cultural algorithm

968 with problem-specific Information. Environmental Modelling and Software, 71, 126–137.

969 Scopus. https://doi.org/10.1016/j.envsoft.2015.06.002

970 78. Liu, X., & Tian, R. (2022). RiverGAN: Fluvial Landform Generation Based on Physical

971 Simulations and Generative Adversarial Network. Journal of Digital Landscape

972 Architecture. https://doi.org/10.14627/537724011

973 79. Lorilla, R. S., Poirazidis, K., Detsis, V., Kalogirou, S., & Chalkias, C. (2020). Socio-

974 ecological determinants of multiple ecosystem services on the Mediterranean landscapes of

975 the Ionian Islands (Greece). Ecological Modelling, 422. Scopus.

976 https://doi.org/10.1016/j.ecolmodel.2020.108994

977 80. Malone, T. W., Rus, D., & Laubacher, R. (2020). Artificial Intelligence and the Future of

978 Work (p. 39) [Research Brief]. Massachusetts Institiute of Technology.

979 https://workofthefuture.mit.edu/wp-content/uploads/2020/12/2020-Research-Brief-Malone-

980 Rus-Laubacher2.pdf

981 81. Mata, J., de Miguel, I., Durán, R. J., Merayo, N., Singh, S. K., Jukan, A., & Chamania, M.

982 (2018). Artificial intelligence (AI) methods in optical networks: A comprehensive survey.

983 Optical Switching and Networking, 28, 43–57. https://doi.org/10.1016/j.osn.2017.12.006

 47

984 82. McCarthy, M. M., & Portner, J. (1980). The Changing Landscape: Communication and

985 Information Technologies. Landscape Architecture, 70(6), 602–611. JSTOR.

986 83. Miller, N. (2017, August 1). Machine Learning with LunchBoxML. Proving Ground.

987 https://provingground.io/2017/08/01/machine-learning-with-lunchboxml/

988 84. Minařík, R., & Langhammer, J. (2016). Use of a Multispectral Uav Photogrammetry for

989 Detection and Tracking of Forest Disturbance Dynamics. ISPRS - International Archives of

990 the Photogrammetry, Remote Sensing and Spatial Information Sciences, 41B8, 711–718.

991 https://doi.org/10.5194/isprs-archives-XLI-B8-711-2016

992 85. Miranda, A., Carrasco, J., González, M., Pais, C., Lara, A., Altamirano, A., Weintraub, A.,

993 & Syphard, A. D. (2020). Evidence-based mapping of the wildland-urban interface to

994 better identify human communities threatened by wildfires. Environmental Research

995 Letters, 15(9). Scopus. https://doi.org/10.1088/1748-9326/ab9be5

996 86. Mirjalili, S., & Dong, J. S. (2020). Multi-objective optimization using artificial intelligence

997 techniques. Springer.

998 87. Mohanty, S. P., Hughes, D. P., & Salathé, M. (2016). Using Deep Learning for Image-

999 Based Plant Disease Detection. Frontiers in Plant Science, 7.

1000 https://doi.org/10.3389/fpls.2016.01419

1001 88. Monge, J. C. (2022, July 13). MidJourney AI Is Now Publicly Accessible—Don’t Miss It.

1002 MLearning.Ai. https://medium.com/mlearning-ai/midjourney-ai-is-now-publicly-

1003 accessible-dont-miss-it-c4c6bb77c375

1004 89. Mouchet, M. A., Lamarque, P., Martín-López, B., Crouzat, E., Gos, P., Byczek, C., &

1005 Lavorel, S. (2014). An interdisciplinary methodological guide for quantifying associations

 48

1006 between ecosystem services. Global Environmental Change, 28(1), 298–308. Scopus.

1007 https://doi.org/10.1016/j.gloenvcha.2014.07.012

1008 90. Naderi, J. R., & Raman, B. (2005). Capturing impressions of pedestrian landscapes used

1009 for healing purposes with decision tree learning. Landscape and Urban Planning, 73(2–3),

1010 155–166. Scopus. https://doi.org/10.1016/j.landurbplan.2004.11.012

1011 91. New Landscape Declaration. (2016). Landscape Architecture Foundation.

1012 https://www.lafoundation.org/take-action/new-landscape-declaration

1013 92. Ngarega, B. K., Masocha, V. F., & Schneider, H. (2021). Forecasting the effects of

1014 bioclimatic characteristics and climate change on the potential distribution of

1015 Colophospermum mopane in southern Africa using Maximum Entropy (Maxent).

1016 Ecological Informatics, 65. Scopus. https://doi.org/10.1016/j.ecoinf.2021.101419

1017 93. Nitoslawski, S. A., Galle, N. J., Van Den Bosch, C. K., & Steenberg, J. W. N. (2019).

1018 Smarter ecosystems for smarter cities? A review of trends, technologies, and turning points

1019 for smart urban forestry. Sustainable Cities and Society, 51, 101770.

1020 https://doi.org/10.1016/j.scs.2019.101770

1021 94. Novikov, A. I., & Ersson, B. T. (2019). Aerial seeding of forests in Russia: A selected

1022 literature analysis. IOP Conference Series: Earth and Environmental Science, 226, 012051.

1023 https://doi.org/10.1088/1755-1315/226/1/012051

1024 95. Ogrin, D. (1994). Landscape architecture and its articulation into landscape planning and

1025 landscape design. Landscape and Urban Planning, 30(3), 131–137.

1026 https://doi.org/10.1016/0169-2046(94)90052-3

1027 96. Peng, J., Zhao, S., Dong, J., Liu, Y., Meersmans, J., Li, H., & Wu, J. (2019). Applying ant

1028 colony algorithm to identify ecological security patterns in megacities. Environmental

 49

1029 Modelling and Software, 117, 214–222. Scopus.

1030 https://doi.org/10.1016/j.envsoft.2019.03.017

1031 97. Petrich, C. (1986). Expert Systems. Landscape Architecture, 76(3), 70–74.

1032 98. Polk, M. (2015). Transdisciplinary co-production: Designing and testing a transdisciplinary

1033 research framework for societal problem solving. Futures, 65, 110–122.

1034 https://doi.org/10.1016/j.futures.2014.11.001

1035 99. Prince, S. J. D. (2012). Computer Vision: Models, Learning, and Inference. Cambridge

1036 University Press.

1037 100. Public Health Agency Canada. (2020, June 4). Challenges and opportunities for public

1038 health made possible by advances in natural language processing, CCDR 46(6) [Research].

1039 https://www.canada.ca/en/public-health/services/reports-publications/canada-

1040 communicable-disease-report-ccdr/monthly-issue/2020-46/issue-6-june-4-2020/natural-

1041 language-processing-subfield-artificial-intelligence.html

1042 101. Queiroz, C., Meacham, M., Richter, K., Norström, A. V., Andersson, E., Norberg, J., &

1043 Peterson, G. (2015). Mapping bundles of ecosystem services reveals distinct types of

1044 multifunctionality within a Swedish landscape. Ambio, 44(1), 89–101. Scopus.

1045 https://doi.org/10.1007/s13280-014-0601-0

1046 102. Raman, T. A., Kollar, J., & Penman, S. (2022). Chapter 17 - SASAKI: Filling the design

1047 gap—Urban impressions with AI. In I. As, P. Basu, & P. Talwar (Eds.), Artificial

1048 Intelligence in Urban Planning and Design (pp. 339–362). Elsevier.

1049 https://doi.org/10.1016/B978-0-12-823941-4.00002-0

1050 103. Rutenbar, R. A. (1989). Simulated annealing algorithms: An overview. IEEE Circuits and

1051 Devices Magazine, 5(1), 19–26. https://doi.org/10.1109/101.17235

 50

1052 104. Sai, M. S., Kumar, K., & Prakash, B. (2020). Design, Analysis and Development of a

1053 Flying Wing UAV for Aerial Seeding and 3D Mapping. In BBVL. Deepak, D. Parhi, & P.

1054 C. Jena (Eds.), Innovative Product Design and Intelligent Manufacturing Systems (pp.

1055 1023–1033). Springer. https://doi.org/10.1007/978-981-15-2696-1_99

1056 105. Schlickman, E. (2020). Assessing automation: Methodological insights from experimenting

1057 with computer vision for public life research. Journal of Landscape Architecture, 15(3),

1058 48–59. https://doi.org/10.1080/18626033.2020.1886515

1059 106. Shanthala Devi, B. S., Murthy, M. S. R., Debnath, B., & Jha, C. S. (2013). Forest patch

1060 connectivity diagnostics and prioritization using graph theory. Ecological Modelling, 251,

1061 279–287. Scopus. https://doi.org/10.1016/j.ecolmodel.2012.12.022

1062 107. Shapira, A., Shoshany, M., & Nir-Goldenberg, S. (2013). Combining analytical hierarchy

1063 process and agglomerative hierarchical clustering in search of expert consensus in green

1064 corridors development management. Environmental Management, 52(1), 123–135. Scopus.

1065 https://doi.org/10.1007/s00267-013-0064-2

1066 108. Slager, C. T. J., & De Vries, B. (2013). Landscape generator: Method to generate landscape

1067 configurations for spatial plan-making. Computers, Environment and Urban Systems, 39,

1068 1–11. Scopus. https://doi.org/10.1016/j.compenvurbsys.2013.01.007

1069 109. Song, Y., Ning, H., Ye, X., Chandana, D., & Wang, S. (2022). Analyze the usage of urban

1070 greenways through social media images and computer vision. Environment and Planning

1071 B: Urban Analytics and City Science, 239980832110646.

1072 https://doi.org/10.1177/23998083211064624
 51

1073 110. Song, Y., Wang, R., Fernandez, J., & Li, D. (2021). Investigating sense of place of the Las

1074 Vegas Strip using online reviews and machine learning approaches. Landscape and Urban

1075 Planning, 205. Scopus. https://doi.org/10.1016/j.landurbplan.2020.103956

1076 111. Souza, J. T. de, Francisco, A. C. de, Piekarski, C. M., & Prado, G. F. do. (2019). Data

1077 Mining and Machine Learning to Promote Smart Cities: A Systematic Review from 2000

1078 to 2018. Sustainability, 11(4), Article 4. https://doi.org/10.3390/su11041077

1079 112. Stamou, Z., Xystrakis, F., & Koutsias, N. (2016). The role of fire as a long-term landscape

1080 modifier: Evidence from long-term fire observations (1922–2000) in Greece. Applied

1081 Geography, 74, 47–55. Scopus. https://doi.org/10.1016/j.apgeog.2016.07.005

1082 113. Suppakittpaisarn, P., Lu, Y., Jiang, B., & Slavenas, M. (2022). How do computers see

1083 landscapes? Comparisons of eye-level greenery assessments between computer and human

1084 perceptions. Landscape and Urban Planning, 227, 104547.

1085 https://doi.org/10.1016/j.landurbplan.2022.104547

1086 114. Sutton, R. S. (1992). Introduction: The Challenge of Reinforcement Learning. In R. S.

1087 Sutton (Ed.), Reinforcement Learning (pp. 1–3). Springer US. https://doi.org/10.1007/978-

1088 1-4615-3618-5_1

1089 115. Szeliski, R. (2010). Computer Vision: Algorithms and Applications. Springer Science &

1090 Business Media.

1091 116. Tack, S. (2021, June 23). NVIDIA Canvas App Launches in Beta. NVIDIA Blog.

1092 https://blogs.nvidia.com/blog/2021/06/23/studio-canvas-app/

1093 117. Talaviya, T., Shah, D., Patel, N., Yagnik, H., & Shah, M. (2020). Implementation of

1094 artificial intelligence in agriculture for optimisation of irrigation and application of

 52

1095 pesticides and herbicides. Artificial Intelligence in Agriculture, 4, 58–73.

1096 https://doi.org/10.1016/j.aiia.2020.04.002

1097 118. Tarca, A. L., Carey, V. J., Chen, X., Romero, R., & Drăghici, S. (2007). Machine Learning

1098 and Its Applications to Biology. PLOS Computational Biology, 3(6), e116.

1099 https://doi.org/10.1371/journal.pcbi.0030116

1100 119. Tebyanian, N. (2020). Application of Machine Learning for Urban Landscape Design: A

1101 Primer for Landscape Architects. https://doi.org/10.14627/537690023

1102 120. Thomson, J., Regan, T. J., Hollings, T., Amos, N., Geary, W. L., Parkes, D., Hauser, C. E.,

1103 & White, M. (2020). Spatial conservation action planning in heterogenous landscapes.

1104 Biological Conservation, 250. Scopus. https://doi.org/10.1016/j.biocon.2020.108735

1105 121. Van Assche, K., Beunen, R., Duineveld, M., & de Jong, H. (2013). Co-evolutions of

1106 planning and design: Risks and benefits of design perspectives in planning systems.

1107 Planning Theory, 12(2), 177–198.

1108 122. van Strien, M. J., & Grêt-Regamey, A. (2022). Unsupervised deep learning of landscape

1109 typologies from remote sensing images and other continuous spatial data. Environmental

1110 Modelling & Software, 155(C). https://doi.org/10.1016/j.envsoft.2022.105462

1111 123. von Haaren, C., Warren-Kretzschmar, B., Milos, C., & Werthmann, C. (2014).

1112 Opportunities for design approaches in landscape planning. Landscape and Urban

1113 Planning, 130, 159–170. https://doi.org/10.1016/j.landurbplan.2014.06.012

1114 124. von Wodtke, M. (1988). Integrating Computer Applications in Higher Education.

1115 Landscape Architecture, 78(5), 90–98. JSTOR.

1116 125. Vovchenko, N., Novikov, A., Sokolov, S., & Tishchenko, E. (2020). A proposed

1117 technology to ensure high-precision aerial seeding of certified seeds. IOP Conference
 53

1118 Series: Earth and Environmental Science, 595, 012066. https://doi.org/10.1088/1755-

1119 1315/595/1/012066

1120 126. Vrontis, D., Christofi, M., Pereira, V., Tarba, S., Makrides, A., & Trichina, E. (2022).

1121 Artificial intelligence, robotics, advanced technologies and human resource management:

1122 A systematic review. The International Journal of Human Resource Management, 33(6),

1123 1237–1266.

1124 127. Wael, S., Elshater, A., & Afifi, S. (2022). Mapping User Experiences around Transit Stops

1125 Using Computer Vision Technology: Action Priorities from Cairo. Sustainability, 14(17),

1126 Article 17. https://doi.org/10.3390/su141711008

1127 128. Wang, X. (2021). Optimization Design of Green Building Landscape Space Environment

1128 Based on VR Virtual Technology. Journal of Physics: Conference Series, 1852(3). Scopus.

1129 https://doi.org/10.1088/1742-6596/1852/3/032035

1130 129. Wang, Z., Zhu, Z., Xu, M., & Qureshi, S. (2021). Fine-grained assessment of greenspace

1131 satisfaction at regional scale using content analysis of social media and machine learning.

1132 Science of the Total Environment, 776. Scopus.

1133 https://doi.org/10.1016/j.scitotenv.2021.145908

1134 130. Wang, Z.-H., Zhao, X., Yang, J., & Song, J. (2016). Cooling and energy saving potentials

1135 of shade trees and urban lawns in a desert city. Applied Energy, 161, 437–444. Scopus.

1136 https://doi.org/10.1016/j.apenergy.2015.10.047

1137 131. Wartmann, F. M., Koblet, O., & Purves, R. S. (2021). Assessing experienced tranquillity

1138 through natural language processing and landscape ecology measures. Landscape Ecology,

1139 36(8), 2347–2365. Scopus. https://doi.org/10.1007/s10980-020-01181-8

 54

1140 132. Wasif, M. (2011). Design and implementation of autonomous Lawn-Mower Robot

1141 controller. 2011 7th International Conference on Emerging Technologies, 1–5.

1142 https://doi.org/10.1109/ICET.2011.6048466

1143 133. Westort, C., & Shen, Z. (2017). Robot in the garden: Preliminary experiments

1144 programming an on-site robot ball assistant to the landscape architect. Journal of Digital

1145 Landscape Architecture, 2017(2), 223–234. Scopus. https://doi.org/10.14627/537629023

1146 134. Westphal, M. I., Field, S. A., & Possingham, H. P. (2007). Optimizing landscape

1147 configuration: A case study of woodland birds in the Mount Lofty Ranges, South Australia.

1148 Landscape and Urban Planning, 81(1–2), 56–66. Scopus.

1149 https://doi.org/10.1016/j.landurbplan.2006.10.015

1150 135. Wu, N., & Silva, E. A. (2010). Artificial Intelligence Solutions for Urban Land Dynamics:

1151 A Review. Journal of Planning Literature, 24(3), 246–265.

1152 https://doi.org/10.1177/0885412210361571

1153 136. Yang, B., & Xu, Y. (2021). Applications of deep-learning approaches in horticultural

1154 research: A review. Horticulture Research, 8, 123. https://doi.org/10.1038/s41438-021-

1155 00560-9

1156 137. Yang, C., & Liu, T. (2022). Social Media Data in Urban Design and Landscape Research:

1157 A Comprehensive Literature Review. Land, 11(10), 1796.

1158 https://doi.org/10.3390/land11101796

1159 138. Yang, J., Fricker, P., & Jung, A. (2022). From Intuition to Reasoning: Analyzing

1160 Correlative Attributes of Walkability in Urban Environments with Machine Learning.

1161 Journal of Digital Landscape Architecture. https://doi.org/10.14627/537724008

 55

1162 139. Yang, L., Iwami, M., Chen, Y., Wu, M., & van Dam, K. H. (2022). Computational

1163 decision-support tools for urban design to improve resilience against COVID-19 and other

1164 infectious diseases: A systematic review. Progress in Planning, 100657.

1165 https://doi.org/10.1016/j.progress.2022.100657

1166 140. Yigitcanlar, T., Desouza, K. C., Butler, L., & Roozkhosh, F. (2020). Contributions and

1167 Risks of Artificial Intelligence (AI) in Building Smarter Cities: Insights from a Systematic

1168 Review of the Literature. Energies, 13(6), Article 6. https://doi.org/10.3390/en13061473

1169 141. Yue, L., Chen, W., Li, X., Zuo, W., & Yin, M. (2019). A survey of sentiment analysis in

1170 social media. Knowledge and Information Systems, 60(2), 617–663.

1171 https://doi.org/10.1007/s10115-018-1236-4

1172 142. Zank, B., Bagstad, K. J., Voigt, B., & Villa, F. (2016). Modeling the effects of urban

1173 expansion on natural capital stocks and ecosystem service flows: A case study in the Puget

1174 Sound, Washington, USA. Landscape and Urban Planning, 149, 31–42. Scopus.

1175 https://doi.org/10.1016/j.landurbplan.2016.01.004

1176 143. Zeiger, M. (2019, February 12). Live and Learn. Landscape Architecture Magazine,

1177 109(12). https://landscapearchitecturemagazine.org/2019/02/12/live-and-learn/

1178 144. Zema, D. A., Lucas-Borja, M. E., Fotia, L., Rosaci, D., Sarnè, G. M. L., & Zimbone, S. M.

1179 (2020). Predicting the hydrological response of a forest after wildfire and soil treatments

1180 using an Artificial Neural Network. Computers and Electronics in Agriculture, 170.

1181 Scopus. https://doi.org/10.1016/j.compag.2020.105280

1182 145. Zhang, L., & Liu, B. (2017). Sentiment Analysis and Opinion Mining. In C. Sammut & G.

1183 I. Webb (Eds.), Encyclopedia of Machine Learning and Data Mining (pp. 1152–1161).

1184 Springer US. https://doi.org/10.1007/978-1-4899-7687-1_907

 56

1185 146. Zhang, R., Zhao, Y., Kong, J., Cheng, C., Liu, X., & Zhang, C. (2021). Intelligent

1186 recognition method of decorative openwork windows with sustainable application for

1187 suzhou traditional private gardens in china. Sustainability (Switzerland), 13(15). Scopus.

1188 https://doi.org/10.3390/su13158439

1189 147. Zhang, Z. (2020). Cybernetic Environment: A Historical Reflection on System, Design,

1190 and Machine Intelligence. 33–40. https://doi.org/10.14627/537690004

1191 148. Zhang, Z., & Bowes, B. (2019). The Future of Artificial Intelligence (AI) and Machine

1192 Learning (ML) in Landscape Design: A Case Study in Coastal Virginia, USA. Journal of

1193 Digital Landscape Architecture, 2–9.

1194 149. Zhu, Y., & Yan, E. (2015). Dynamic subfield analysis of disciplines: An examination of

1195 the trading impact and knowledge diffusion patterns of computer science. Scientometrics,

1196 104(1), 335–359. https://doi.org/10.1007/s11192-015-1594-6

1197