criado em:

• 08-05-2025
• 16:57 relacionados:
• notas:

1. Deep Learning - uma abordagem interdisciplinar

• tags: #DR
• Fontes & Links:

The Entangled Geographies of Artificial Intelligence: Spatial, Territorial, and Environmental Dimensions

1. Introduction

Artificial Intelligence (AI) is rapidly transcending its status as a purely technological advancement to become a profoundly geographical phenomenon, one with far-reaching spatial, territorial, and environmental implications. The swift and pervasive proliferation of AI technologies across myriad sectors of society 1 necessitates an urgent and critical geographical analysis. Such an analysis is essential to unravel the complexities of AI’s uneven development, the inequitable distribution of its benefits and burdens, and its transformative, often disruptive, effects on both physical landscapes and human societies. The core research problem this report addresses is the multifaceted manner in which AI shapes, and is shaped by, geographical contexts, frequently exacerbating existing socio-spatial inequalities or forging new ones.4 The burgeoning field of AI presents a complex array of questions that demand rigorous geographical inquiry.1 While AI’s proliferation has been swift 3, the systematic engagement of geography as a discipline appears, in part, as a critical response to technological trajectories already in motion. Much like other academic fields grappling with AI’s societal integration, geography now confronts the task of analyzing systems and impacts that have developed with considerable momentum, often outpacing deep, critical geographical scrutiny in their foundational stages.8 This context underscores the urgency of applying geographical lenses to understand how AI technologies are not only shaping, but are also fundamentally shaped by, diverse geographical contexts.

The significance of deploying geographical perspectives to scrutinize AI cannot be overstated. Geography, with its intrinsic focus on space, place, scale, environment, and territory, offers a unique and indispensable toolkit for dissecting the complex societal integration of AI.1 Geographical inquiry possesses the distinct advantage of being able to reveal the often-hidden materialities, entrenched power dynamics, and profound socio-environmental consequences that are frequently obscured in purely technical or economistic discourses surrounding AI.12 By illuminating spatial injustices and informing the development of alternative, place-based approaches to AI, geography can contribute significantly to fostering more just, equitable, and sustainable AI futures.14 The very premise of investigating AI’s spatial, territorial, and environmental dimensions positions it not merely as an inert tool, but as an active agent in the reconfiguration of socio-spatial and socio-environmental relations. Its deployment is actively transforming how societies are organized in space, how territories are governed, and how environments are impacted 12, necessitating an analytical approach that recognizes AI as a dynamic force with the capacity to produce new spatial realities and power geometries.

This report aims to conduct a comprehensive investigation into how the discipline of geography analyzes AI through its spatial, territorial, and environmental dimensions, addressing seven core research questions. The analysis will integrate established theoretical frameworks from geography and critical social theory with contemporary AI developments. It will draw upon empirical evidence from a diverse range of case studies, encompassing the Global North and South, as well as urban and rural contexts, and will incorporate critical perspectives from postcolonial, feminist, and Indigenous scholarship. The report is structured as follows: Section 2 develops a coherent Theoretical Framework. Section 3 outlines the Methodological Approach. Section 4 provides an in-depth Analysis of Key Dimensions, addressing each research question. Section 5 offers a SWOT Analysis of AI from a geographical standpoint. Finally, Section 6 presents the Conclusion and suggests Future Directions for research.

2. Theoretical Framework

To comprehensively analyze the multifaceted geographical dimensions of Artificial Intelligence, this report integrates several key theoretical frameworks. These frameworks, drawn from geography and critical social theory, provide robust lenses through which to examine AI’s development, deployment, and societal implications. The convergence of critique from such a diverse range of established theories underscores AI’s profound and multifaceted impact as a socio-spatial ordering force; it is not merely a technical object but a site where multiple forms of power, inequality, and environmental relation are being contested and reconfigured.

Spatial Justice: Coined by scholars like Edward Soja and building on the work of David Harvey, spatial justice refers to the fair and equitable distribution in space of socially valued resources, opportunities, and burdens.17 AI development, access, and its myriad impacts can significantly create or exacerbate spatial injustices. For instance, the intense geographical concentration of AI research, investment, and benefits in a few global hubs 9 contrasts sharply with the externalization of environmental harms and exploitative labor to peripheral regions.5 This creates “locational discrimination” 17, where certain populations are disproportionately disadvantaged due_to_their geographical location, a situation potentially amplified by AI systems trained on biased datasets or deployed without sensitivity to local socio-spatial contexts.4

Political Ecology: This approach, advanced by thinkers like Piers Blaikie, Raymond Bryant, and Paul Robbins, examines the complex relationships between political, economic, and social factors, and environmental issues and changes. Applied to AI, political ecology illuminates the power dynamics inherent in the extraction of resources for AI hardware (e.g., rare earth elements, cobalt, lithium) 12, the significant environmental burden of data centers (energy and water consumption, land use) 1, and the political ramifications of employing AI in natural resource management and environmental governance.22 Central to this analysis is the concept of the “eco-political economy of AI” 12, which meticulously links AI’s entire production chain—from mineral extraction to e-waste—to patterns of environmental degradation and social inequality.

Actor-Network Theory (ANT): Developed by Bruno Latour, Michel Callon, and John Law, ANT emphasizes the role of both human and non-human actors (including technologies) in the co-construction of socio-technical networks.25 From an ANT perspective, AI systems are viewed as complex assemblages comprising algorithms, data, hardware, human developers, users, institutions, and regulatory frameworks, all of which mutually shape one another’s actions and characteristics. ANT is particularly useful for revealing the subtle and explicit power dynamics within AI development and deployment, where algorithms themselves can exert a form of agency and influence outcomes.25 The ANT concept of “translation”—how actors (including AI) interpret and enroll others into their projects—helps understand how AI systems make sense of and act upon the world, sometimes with unforeseen or undesirable consequences.25

Critical Data Studies: Scholars such as Nick Couldry, Ulises Mejias, and Rob Kitchin have pioneered Critical Data Studies, which interrogates the social, legal, political, and economic dimensions of datafication and AI.27 A key concept here is “data colonialism” 13, describing how the extraction of data from populations, particularly in the Global South, often mirrors historical colonial exploitation, with value accruing to powerful entities predominantly in the Global North. AI systems, built upon these vast datasets, can inadvertently embed, amplify, and automate biases present in the data, leading to discriminatory outcomes and the reinforcement of existing power structures and inequalities.4

Postcolonial and Decolonial Theories: Drawing from thinkers like Edward Said, Walter Mignolo, and Achille Mbembe, postcolonial theory critiques the enduring legacies of colonialism. In the context of AI, this lens is applied to analyze phenomena such as “robotic colonisation” 30 or “AI colonialism”.5 These terms describe how AI technologies originating from developed nations can extend influence and control over less-developed regions, thereby impacting their sovereignty, creating technological dependencies, and perpetuating unequal power relations.31 Historical colonial patterns of resource extraction and labor exploitation are often mirrored in AI’s global supply chains (e.g., mineral extraction, data labeling) and data practices.5 Decolonial approaches call for challenging these dynamics and valuing alternative knowledge systems, including Indigenous epistemologies, in AI development.15

Feminist Technology Studies: Building on the work of Donna Haraway, Judy Wajcman, and Catherine D’Ignazio & Lauren Klein, feminist technology studies offer critical perspectives on how gender, power, and technology intersect. This framework highlights how AI systems can embed and perpetuate gender biases, often stemming from unrepresentative training data or the perspectives of predominantly male development teams.37 The gendered division of labor within the AI industry and the underrepresentation of women and marginalized genders can lead to technologies that fail to serve diverse needs or actively discriminate.7 Consequently, feminist AI design principles advocate for approaches that promote equity, inclusion, intersectionality, and challenge patriarchal power structures embedded within technological systems.37

Across these diverse theoretical perspectives, a recurring and crucial interconnection emerges: the triad of AI’s material infrastructure, the power relations it engenders or reinforces, and the knowledge systems (epistemologies) it privileges or marginalizes. Political ecology and analyses of resource extraction directly link AI’s physical hardware to environmental degradation and exploitative labor, which are inherently issues of power.12 Critical data studies and postcolonial theories reveal how data, an immaterial yet potent resource, is extracted under specific power relations, often guided by dominant Western epistemologies, to the detriment of the Global South.13 Feminist theories demonstrate how biased datasets (reflecting particular knowledge and power structures) lead to discriminatory AI tools (materiality) that reinforce gendered power imbalances.37 Critiques from Indigenous epistemologies argue that AI’s dominant Western knowledge framework clashes with holistic, contextual Indigenous knowledge, profoundly impacting how AI is developed, who benefits, and whose power is amplified or diminished.15 Therefore, a geographical analysis of AI must consider these three elements—materiality, power, and knowledge—not in isolation, but as a co-constitutive triad, where changes in one element can reconfigure the others, potentially paving the way for more equitable and sustainable AI geographies.

3. Methodological Approach

This research report employs a qualitative, critical-interpretive methodology, primarily based on a comprehensive synthesis of existing peer-reviewed academic literature, supplemented by authoritative reports from research institutions, non-governmental organizations (NGOs), and international bodies. The approach is multi-scalar, examining AI geographies from the global level, such as international power relations shaped by AI dominance 18, down to the local level, including urban transformations driven by smart city initiatives 16 and the emergence of community-based AI applications.43

The primary data sources consist of scholarly articles drawn from a wide array of disciplines, including geography, science and technology studies (STS), urban planning, environmental studies, critical data studies, and related interdisciplinary fields. This diverse body of literature is exemplified by sources covering AI sustainability 1, GeoAI applications 10, political ecology of AI 12, actor-network perspectives 25, critical data and postcolonial critiques 27, feminist analyses 37, and Indigenous epistemologies in AI.15 Reports from entities such as the Stanford AI Index 44, the United Nations 6, UNESCO 7, and Oxford Insights (AI Readiness Index) 45, as well as investigative reports like those from Global Witness on resource extraction 19, provide crucial empirical data and contextual information. This includes quantitative data on AI investment trends 44, the state of digital infrastructure globally 31, documented environmental impacts of AI systems 1, and disparities in technological access.4 Qualitative data is drawn from numerous case studies embedded within these sources, which examine specific AI deployments and their geographical consequences, such as smart city projects in diverse contexts 42, AI initiatives in Global South healthcare 53, and Indigenous AI projects.36 The complexity of AI’s geographical dimensions necessitates a methodological sensibility that can synthesize insights from various research traditions. The sources drawn upon for this report themselves employ a diverse array of methods, including quantitative surveys 38, geospatial analysis 54, ethnographic fieldwork 56, and bibliometric reviews 10, underscoring the need for a pluralistic approach to understanding this multifaceted phenomenon.

The analytical techniques employed include:

1. Thematic Analysis: Identifying, analyzing, and reporting recurring themes, patterns, contradictions, and critical arguments across the diverse body of literature and case studies pertinent to each of the report’s research questions.
2. Comparative Analysis: Systematically comparing AI development trajectories, socio-environmental impacts, and governance approaches across different geographical contexts. This includes comparisons between the Global North and Global South, urban and rural settings, and specific national or regional case studies (e.g., AI in Brazil 56, data center development in Singapore 60, algorithmic governance in China 61).
3. Critical Discourse Analysis: Examining the dominant narratives, metaphors, and framings used in discussions about AI, particularly concerning concepts like “development,” “progress,” “efficiency,” and “risk.” This involves analyzing how these discourses shape policy, public perception, and ultimately, geographical outcomes.
4. Integration of Theoretical Frameworks: Consistently applying the conceptual tools from the established theoretical framework—spatial justice, political ecology, Actor-Network Theory, critical data studies, postcolonial theory, and feminist technology studies—to interpret empirical findings and case study evidence. This allows for a theoretically informed and critically nuanced understanding of AI’s geographical dimensions.

The emphasis on incorporating multiple case studies from diverse geographical contexts is central to this methodology. Abstract theories and global trends regarding AI, such as “data colonialism” or “spatial justice,” gain concrete meaning and are tested through their application to specific, place-based realities. Case studies—whether examining AI in Brazilian tech work 56, AI for prenatal care in Guatemala 53, or data center impacts in Singapore 60—provide the essential empirical texture needed to understand how global AI processes interact with local cultures, power structures, and environmental conditions. Without this grounding in specific instances, the analysis risks becoming overly generalized and detached from the lived experiences and material realities that define AI’s geographies.

4. Analysis of Key Dimensions

This section delves into the core research questions, analyzing the multifaceted geographical dimensions of Artificial Intelligence. Each subsection addresses a specific aspect, drawing upon the theoretical framework and methodological approach outlined previously.

4.1 Geopolitical Dimensions of AI Development

The development of Artificial Intelligence is not occurring in a geographical vacuum; rather, it is characterized by intense geographical concentration, which profoundly shapes international power relations and poses significant challenges to digital sovereignty and technological autonomy, particularly for peripheral regions.

Geographical Concentration and Power Relations: AI development is overwhelmingly concentrated in a few global hubs, most notably Silicon Valley in the United States and key cities in China such as Beijing and Shenzhen.9 These centers, along with other significant nodes in Europe and parts of Asia, attract the lion’s share of global AI investment, talent, and patent registrations.18 For instance, in 2024, corporate investment in AI rebounded significantly, with generative AI alone attracting $33.9 billion globally, and U.S. institutions producing a markedly higher number of notable AI models compared to other regions, though Chinese models are rapidly closing the quality gap.44 This concentration fosters a “multitiered global AI ecosystem” 18, where a small cadre of nations and corporations lead innovation and value capture, a larger group strives to keep pace, and a substantial portion of the world risks being left further behind. Such consolidation of AI capabilities directly translates into enhanced technological dominance for these leading nations, granting them unprecedented influence over AI’s future trajectory, including the shaping of technical standards, ethical frameworks, and global governance mechanisms.18 The rivalry between the U.S. and China in AI, for example, is not merely a bilateral competition but a defining feature of the current geopolitical landscape, compelling other nations to navigate this complex dynamic.41

Digital Sovereignty and Technological Autonomy in Peripheral Regions: For nations in the Global South and other peripheral regions, this concentration poses acute challenges to their digital sovereignty and technological autonomy.31 Digital sovereignty, the ability of a state to govern its own digital destiny, is threatened by reliance on foreign-developed AI systems for critical infrastructure, financial services, and even national defense, potentially undermining national security and policy independence.18 African states, for example, are actively strategizing how to navigate AI sovereignty amidst intense global geopolitical competition and the pervasive influence of major global players (U.S., EU, China) on the development of AI frameworks within the continent.64 Brazil presents a complex case, acting as a hub for AI development linked to extractive industries, which challenges simplistic narratives of the country as merely a “data colony” but simultaneously highlights its subordinate integration into the global AI market.56 The overarching risk is the emergence of an “AI Iron Curtain,” as described by some analysts 65, which could split the world into distinct geopolitical blocs dominated by Big Tech powers, further marginalizing the digital sovereignty of the majority world. This dynamic is not static but involves a self-reinforcing cycle: initial advantages in expertise, data access, and capital in dominant AI hubs attract further talent and investment, solidifying their leadership and making it progressively harder for peripheral regions to develop independent capabilities and break free from technological dependency.18

Spatial Inequality in AI Access and Development: Patterns of spatial inequality are starkly evident in AI access and development. Quantitative data, such as the Oxford Insights Government AI Readiness Index, reveals significant disparities both between and within global regions.45 Sub-Saharan Africa, for instance, scores lowest on average, though countries like Mauritius and South Africa show higher readiness.45 Qualitative evidence further underscores these inequalities.4 This “digital divide” 4 is characterized by a lack of robust digital infrastructure (internet connectivity, cloud computing, reliable energy), insufficient AI skills and literacy, and limited institutional capacity in many developing countries, hindering their ability to harness AI’s benefits and innovate locally.18 Developing countries, despite contributing essential inputs like raw data or low-wage labor to the AI value chain, often reap minimal benefits from high-value outputs, which remain concentrated in developed economies.18

Digital sovereignty for these peripheral regions is thus a deeply contested terrain. It involves more than just data localization policies or attempts to build local infrastructure; it fundamentally requires navigating complex power dynamics in a global system dominated by a few AI superpowers and their associated corporations.31 This entails careful negotiation of foreign investment terms, active participation in shaping global AI governance frameworks, and resisting the undue influence of external actors who may not prioritize local needs or equitable development. The “tension between foreign investment and local interests” 31 encapsulates this ongoing struggle, where the drive for technological advancement must be balanced against the imperative to maintain national autonomy and ensure that AI development serves broad societal well-being rather than narrow corporate or geopolitical interests.

Table 1: Concentration of AI Development and Peripheral Region Challenges

Note: Investment figures are illustrative of global trends; specific hub investments vary. AI Readiness Index scores are examples and vary widely within regions.

4.2 Environmental Geography of AI Infrastructure

The material infrastructure underpinning Artificial Intelligence—comprising data centers, vast networks, energy systems, and the full lifecycle of hardware components—is profoundly transforming physical landscapes and exerting considerable, often geographically uneven, environmental pressures.

Material Infrastructure and Landscape Transformation: The physical backbone of AI includes energy-intensive data centers, extensive terrestrial and submarine fiber optic networks for data transmission, and the complex global supply chains involved in manufacturing AI hardware, from resource extraction to electronic waste (e-waste) disposal.1 Data centers are particularly land-intensive 60, often sited in areas offering cheap land and energy, which can include peri-urban or rural locations, sometimes with significant ecological opportunity costs.1 Case studies, such as the development of “data center alley” in Loudoun County, Virginia, which saw a nearly 65% increase in water consumption between 2019 and 2023 due to AI computing growth 49, or Singapore’s challenges as a digital hub with physical land and resource limitations 60, illustrate this direct land transformation and resource strain. Remote sensing and GIS, increasingly augmented by AI, are used to monitor such urban growth and land use/land cover (LULC) changes 55, although specific quantification of land transformed solely for AI data centers remains an area for more detailed research.55 The installation of terrestrial and submarine fiber optic cables also causes localized habitat disruption, particularly in sensitive marine environments during the laying process.68

Mapping and Measuring Environmental Impacts:

• Energy Consumption: AI model training and inference, along with the 24/7 operation of data centers, demand massive amounts of electricity.1 Globally, data center electricity consumption was estimated at 460 terawatt-hours in 2022 and is projected to approach 1,050 TWh by 2026, partly driven by generative AI.21 The environmental impact varies geographically, depending on the local energy mix (reliance on fossil fuels versus renewables) and the resulting carbon intensity.1
• Water Usage: Data centers consume substantial volumes of water for cooling their hardware.1 Microsoft reported consuming 6.4 million cubic meters of water in 2022, and Google 19.5 million cubic meters.49 This is particularly concerning in arid or water-stressed regions, where data centers compete with agricultural and municipal water needs, potentially leading to aquifer depletion and social conflict.1
• Carbon Emissions: The significant energy consumption directly translates into greenhouse gas emissions, especially when sourced from fossil fuels.1 Tools like CodeCarbon, Eco2AI, and Carbontracker have been developed to help estimate and report these emissions.1
• Resource Extraction & E-waste: The production of AI hardware necessitates the extraction of rare earth elements and other minerals (detailed further in section 4.5), and the rapid obsolescence of AI hardware contributes to a growing global e-waste problem.12

The promotion of AI and digital technologies as solutions for environmental sustainability—so-called “digital green” initiatives 60—thus encounters a critical paradox: the very tools intended for decarbonization and efficiency are themselves highly resource-intensive and carry a substantial, growing material and environmental footprint.48 This necessitates a critical geographical perspective that deconstructs simplistic “green AI” narratives and demands a comprehensive lifecycle assessment of AI’s environmental costs against its purported benefits.

Furthermore, these environmental impacts are not distributed evenly. A consistent pattern emerges where the more polluting and resource-intensive aspects of the AI lifecycle—such as mineral extraction, manufacturing in regions with less stringent environmental regulations, the siting of data centers reliant on cheap (often fossil-fuel based) energy, and e-waste disposal—tend to be concentrated in the Global South or economically marginalized communities within the Global North.5 Meanwhile, the primary economic benefits, intellectual property, and control over AI technologies accrue predominantly to corporations and nations in the Global North. This geographical burden-shifting means AI’s environmental footprint is deeply interwoven with issues of environmental justice and neocolonial dynamics.

Mitigation Strategies: Addressing AI’s environmental footprint requires multi-scalar strategies. These include developing more energy-efficient AI models and hardware (e.g., sparse models, quantization, specialized AI processors like NVIDIA’s Blackwell GPUs) 77, designing sustainable data centers (powered by renewable energy, employing advanced water-efficient cooling technologies) 77, implementing circular economy principles for hardware (recycling, extending product lifecycles) 77, and enacting robust policy interventions, such as moratoria or caps on data center energy demand in resource-constrained areas.78 Industry initiatives like the Climate Neutral Data Centre Pact in Europe, which aims for climate-neutral operations by 2030 and has reportedly seen participating facilities reduce energy consumption by 25% since 2022 77, and public-private partnerships like those between the U.S. government and tech leaders to advance green data center technologies 77, offer examples of governance efforts. Ironically, AI itself can be leveraged for better environmental monitoring and resource optimization 77, but this potential must be weighed against its own footprint.

Table 2: Comparative Environmental Footprint of AI’s Material Infrastructure

4.3 AI and Urban Transformation

Artificial Intelligence is rapidly and profoundly reconfiguring urban spaces and governance structures, moving beyond earlier “smart city” paradigms to usher in an era of “AI urbanism” or the “AI city”.16 Unlike the vision of the smart city as a seamlessly integrated and steerable machine, the AI city is characterized by systems with greater autonomy, anticipatory capabilities, and emergent logics, making it inherently more unpredictable and adaptive.16 This transformation necessitates new theoretical frameworks to understand the evolving relationships between algorithmic systems and territorial organization.

Theoretical Lenses for Urban AI:

• Assemblage Theory: This perspective is crucial for understanding AI urbanism, as it views the city as a complex interplay of human and non-human elements—including AI technologies, existing physical infrastructure, social practices, data flows, and governance frameworks—that mutually constitute and reconfigure urban realities.16 AI is not merely imposed upon cities; it becomes an active component of the urban assemblage, changing and being changed by its context.
• Algorithmic Governance and Algorhythmic Governance: The concept of algorithmic governance describes how automated digital systems increasingly regulate societal functions.61 Within urban contexts, this manifests in AI-driven systems managing everything from traffic flow 81 and resource allocation (e.g., water, energy) 86 to public safety and service delivery.85 Coletta and Kitchin’s notion of “algorhythmic governance” 84 specifically highlights how city-scale Internet of Things (IoT) infrastructures, coupled with AI, are used to measure, monitor, and actively regulate the polymorphic temporal rhythms of urban life, effectively shaping the “heartbeat” of the city.
• Platform Urbanism: This framework examines how digital platforms, often powered by AI, mediate a growing range of urban services and interactions (e.g., mobility, housing, commerce), raising critical questions about data ownership, market power, labor conditions, and the public accountability of these privately-owned urban infrastructures.61

Smart City Initiatives and Algorithmic Governance: Smart city initiatives worldwide are increasingly incorporating AI. Examples include AI-powered waste management in Seoul, where bins identify and sort waste and optimize collection routes 85; smart traffic management systems in Singapore, Dubai, and Indian cities that use AI sensors to adjust signals, predict congestion, and coordinate public transport 42; and AI-driven urban planning tools like the Open Building Insights (OBI) platform used in Africa and India to model urban growth and infrastructure needs.42 However, the implementation and outcomes of such initiatives vary significantly, particularly between the Global North and South. Cities in the Global South often face substantial barriers, including infrastructural deficits (broadband, energy), a persistent digital divide, inadequate policy and governance frameworks, and financial constraints, which can limit the effectiveness and equitable distribution of AI’s benefits.52 China’s health code infrastructure, an example of large-scale algorithmic governance, demonstrates AI’s capacity to control urban mobility but also highlights vulnerabilities to system bugs and potential for social exclusion.61 The evolution from “smart city” to “AI city” is marked by AI’s capacity for autonomous action, introducing new uncertainties. Urban governance must adapt from mere management to a form of co-existence with these increasingly agential AI systems, requiring new forms of oversight and public accountability.

Spatial Surveillance and Urban Morphology: The proliferation of AI-powered spatial surveillance systems, including facial recognition technology for law enforcement and predictive policing algorithms 85, is reshaping the experience and governance of public space. While often framed in terms of enhancing safety and efficiency, these systems raise profound ethical concerns regarding privacy, algorithmic bias (e.g., racial profiling), data security, and the potential for a chilling effect on civil liberties.85 The physical morphology of cities is also being influenced by AI, from AI-assisted architectural design optimizing building layouts and energy performance 88 to AI tools analyzing urban vitality based on diverse datasets like heat maps and social media check-ins to inform urban renewal.88

The deployment of AI in urban governance directly intersects with concerns of spatial justice. If AI systems for resource allocation, service delivery, or policing are trained on biased data or implemented without considering existing inequalities, they can reinforce or create new forms of locational discrimination and social stratification within cities.14 For example, if an AI system optimizing public transport routes underrepresents certain neighborhoods due to data gaps, those areas may receive poorer service, thereby deepening spatial injustice.4 Achieving spatially just urban futures in the age of AI thus requires critical attention to the design, data inputs, deployment contexts, and governance of urban AI systems, alongside robust mechanisms for public participation, contestation, and redress for algorithmically-generated inequalities.

Table 3: AI in Urban Contexts: Comparative Case Studies

4.4 Cultural Geography and Alternative AI Paradigms

The dominant paradigms of Artificial Intelligence development, largely rooted in Western epistemologies and cultural norms 7, exhibit homogenizing tendencies that often fail to adequately represent, serve, or respect the diverse cultural landscapes of the world.57 This can lead to the perpetuation of “cognitive imperialism” 35 or “data colonialism” 13, where one set of cultural assumptions is implicitly or explicitly imposed through technology. A critical cultural geography of AI, therefore, seeks to challenge these tendencies by advocating for culturally-situated, place-based approaches and by exploring how alternative epistemologies, particularly Indigenous ways of knowing, can inform more diverse and equitable AI geographies.

Countering Homogenization with Place-Based AI: The development of AI that is sensitive to local contexts, values, and needs is paramount.36 As research by the Ada Lovelace Institute suggests, “place matters” significantly in how different communities perceive AI’s opportunities and risks, and how they wish to participate in its governance.43 This calls for evaluation methods that go beyond simplistic metrics of accuracy or efficiency to incorporate “thick evaluations” of cultural representation, steeped in communities’ own understanding of what constitutes appropriate and respectful portrayal.57 Case studies from the Global South offer compelling examples of such locally adapted AI. For instance, the AymurAI project in Latin America co-developed a feminist AI tool to help address data gaps on gender-based violence, grounding its design in local feminist activism and judicial needs.40 In Ghana, an AI solution was developed to improve sexual and reproductive health (SRH) information access for adolescents with disabilities, tailored to their specific communication needs.53 Similarly, an AI chatbot in Turkey provides culturally sensitive SRH advice to refugee women, overcoming language and social barriers.53 These projects demonstrate a commitment to designing AI with and for specific communities, rather than imposing generic solutions.

Indigenous Epistemologies and Alternative AI Geographies: A significant conflict exists between the decontextualized, often reductionist, knowledge frameworks embedded in mainstream AI and the holistic, contextual, relational, and community-based epistemologies characteristic of many Indigenous cultures.15 Recognizing this “epistemological clash” is the first step towards envisioning alternative AI geographies. Frameworks are emerging that advocate for infusing AI development with Indigenous epistemologies, not merely as a token gesture, but as a fundamental reorientation of AI’s design principles to foster genuine inclusivity, respect diverse knowledge systems, and challenge colonial legacies in technology.35 Examples include the development of a Kaupapa Māori model for creating Māori Information Technology artefacts in Aotearoa/New Zealand, which explicitly grounds IT development in Māori worldviews and values 36, and initiatives focused on using AI for Indigenous language vitalization in ways that respect community ownership and protocols.36 Such approaches move beyond merely correcting bias in existing AI models to co-creating AI systems that are aligned with Indigenous self-determination and cultural continuity. The failure of mainstream AI to inclusively represent different cultures 57 is a direct consequence of its epistemological narrowness, underscoring that technical fixes for bias are insufficient without a deeper engagement with and respect for non-Western knowledge systems.

Locally Adapted Applications and Resistance: Beyond Indigenous contexts, numerous AI applications are being tailored to specific local needs in the Global South, particularly in agriculture (e.g., Microsoft’s AI-driven sowing app in India, AI-powered plant disease diagnosis in Uganda 90), healthcare (e.g., AI for prenatal ultrasound interpretation by midwives in Guatemala, AI-assisted depression detection in Bangladesh 53), and education. A key consideration for such applications is the need for offline functionality and localized training data to ensure relevance and accessibility in resource-constrained environments.90 Alongside these adaptive efforts, there is also active community and artistic resistance to the imposition of dominant, potentially colonizing, AI paradigms.29 Initiatives like Te Hiku Media in Aotearoa/New Zealand, which works on Māori language technologies, the development of tools like Glaze to protect artists from AI style mimicry 91, and broader data leverage strategies where communities withhold or “poison” data to disrupt harmful AI models 91, represent forms of “technologies of resistance.” This resistance is not merely oppositional; it is often a generative force, prompting the development of alternative AI tools, practices, and governance models that better serve local needs and values, thereby shaping alternative AI geographies from the ground up. These efforts often align with feminist AI design principles that critically examine power and advocate for technology built in service of social justice and equity.37

Table 4: Place-Based and Indigenous AI: Examples

4.5 Resource Extraction and Environmental Justice

The gleaming edifice of Artificial Intelligence, often perceived as an immaterial force of the digital realm, rests upon a profoundly material foundation rooted in the intensive extraction of natural resources. This “eco-political economy of AI” 12 begins with the mining of critical minerals—including rare earth elements (REEs), cobalt, lithium, copper, and nickel—essential for manufacturing the specialized hardware (CPUs, GPUs, TPUs, memory, and circuitry) that powers AI systems, as well as significant water resources consumed throughout the hardware lifecycle.12 This extractive imperative creates new geographies of environmental injustice, disproportionately burdening communities and ecosystems in the Global South.

Global Supply Chains and Environmental Degradation: The global supply chains for AI hardware are complex and geographically dispersed. Mining for key minerals is often concentrated in specific regions of the Global South: over 70% of the world’s cobalt is sourced from the Democratic Republic of Congo (DRC) 92; South America’s “Lithium Triangle” (Argentina, Bolivia, Chile) is a major producer of lithium 92; and China dominates the mining and processing of REEs, sometimes through operations in other countries like Myanmar under questionable ethical and environmental standards.19 These raw materials are then typically shipped for processing and manufacturing, often to East Asia and other industrial hubs, before being assembled into final products and distributed globally.12

The environmental consequences at extraction sites are severe. Cobalt mining in the DRC, for instance, is linked to extensive deforestation as forests are cleared for mines, widespread water contamination from mining effluents that kill aquatic life and poison drinking water sources, and soil pollution.5 Similarly, lithium extraction in the arid salt flats of South America consumes vast quantities of water, depleting scarce local water resources crucial for local communities and fragile ecosystems.92 The mining of REEs often involves hazardous chemicals and generates significant waste.12 Beyond extraction, the manufacturing of AI hardware, particularly semiconductors, is also highly water-intensive and can lead to chemical pollution if not managed stringently, posing risks to communities near electronics manufacturing hubs, many of which are located in the Global South or in less regulated zones.75 This “clean tech” paradox, where AI’s digital promise is built on a foundation of “dirty” material extraction and processing, is a critical geographical concern.

Social Implications and Environmental Injustice: The social impacts of this extractive geography are equally devastating and unjust. Reports consistently document severe human rights abuses in mining operations, including the use of child labor (especially in artisanal cobalt mines in the DRC), perilous and unsafe working conditions leading to injuries and fatalities, and the forced displacement of Indigenous peoples and local communities from their ancestral lands to make way for mining projects.5 Workers and surrounding communities often suffer severe health impacts from exposure to toxic dust, heavy metals, and contaminated water, including respiratory diseases and birth defects.92 These extractive activities lead to the loss of traditional livelihoods, cultural heritage, and food security for affected populations.12

These spatial patterns of resource extraction and burden-shifting are not accidental but are deeply structured by global economic inequalities and historical power dynamics, often mirroring and reinforcing colonial relationships where the resources of the Global South are exploited for the technological and economic advancement of the Global North.5 Environmental injustice is thus not merely a byproduct but a constitutive feature of AI’s current material geography. AI-driven environmental justice tools themselves can be used to map pollution hotspots and identify communities disproportionately affected by environmental hazards 97, potentially including those generated by AI’s own supply chain. Addressing AI’s environmental impact, therefore, requires more than technological fixes; it demands a fundamental restructuring of global supply chains, robust enforcement of environmental and labor standards, and a commitment to environmental justice that challenges these deeply embedded extractive relationships and ensures that the costs of AI are not disproportionately borne by the most vulnerable.

Table 5: Extractive Geographies of AI: Critical Resources and Injustices

4.6 AI Applications in Geographic Information Systems

The integration of Artificial Intelligence, particularly machine learning (ML) and deep learning (DL) techniques, with Geographic Information Systems (GIS) is giving rise to a powerful subfield known as GeoAI.2 This synergy is transforming environmental monitoring and governance by significantly enhancing the capabilities of GIS to process and analyze vast and complex geospatial datasets, automate analytical workflows, improve predictive modeling, and optimize operational decision-making.58

Specific Applications in Environmental Monitoring and Governance:

• Climate Change Monitoring and Adaptation: AI-enhanced GIS (AI-GIS) is instrumental in analyzing historical and real-time climate data, modeling the impacts of climate change (e.g., sea-level rise, extreme weather events), predicting land use/land cover (LULC) changes, and supporting adaptation strategies.8 For instance, AI algorithms can analyze satellite imagery and meteorological data to predict air quality levels with high accuracy 101 or model the formation and intensity of urban heat islands based on LULC transformations.55
• Biodiversity Conservation: AI is revolutionizing biodiversity research by enabling automated species identification from images (e.g., camera traps) and acoustic recordings, developing predictive habitat models, mapping habitat distribution and connectivity, tracking wildlife populations and migration patterns, and assessing ecological risks.101 Examples include AI applications for monitoring aquatic biodiversity by analyzing environmental DNA or hydroacoustic data 104, and using AI with remote sensing for tracking urban wildlife and managing invasive species with accuracies often exceeding 90%.105
• Disaster Management (Preparedness, Response, Resilience): AI-GIS plays a critical role across all phases of disaster management. It is used for comprehensive risk assessment (e.g., identifying vulnerable populations and infrastructure), developing early warning systems for hazards like floods and landslides, real-time monitoring of disaster progression using satellite or drone imagery, optimizing evacuation routes, and coordinating resource allocation during emergency response.24
• Natural Resource Management: Beyond specific crises, AI-GIS supports broader natural resource management, including sustainable forest management (e.g., detecting illegal logging, monitoring forest health 98), optimizing water resource allocation and monitoring water quality 101, and enhancing precision agriculture to improve yields while minimizing environmental impact.54

Political Implications of Algorithmic Approaches: The increasing reliance on AI-GIS in environmental governance is not without significant political implications. While these technologies offer unprecedented capabilities, they also introduce new complexities related to power, equity, and accountability, functioning as a double-edged sword.

• Power Dynamics and Data Control: A critical question concerns who controls the AI-GIS tools, the underlying data, and the knowledge generated from them. There is a potential for these powerful technologies to reinforce existing power structures or create new “data elites,” where access to and control over information translates into influence over environmental decision-making.23
• Equity and Access: Significant disparities in access to AI-GIS technology, data, and the necessary expertise can lead to an unequal distribution of benefits from improved environmental governance.108 If local communities, Indigenous groups, or less-resourced nations cannot meaningfully participate in or access these systems, they risk being further marginalized or becoming mere subjects of top-down monitoring and management, rather than active participants in stewarding their own environments.103
• Transparency and Accountability: The “black box” nature of some complex AI algorithms poses challenges for transparency and interpretability in decision-making processes.24 If the reasoning behind an AI-driven environmental policy or resource allocation decision is opaque, it becomes difficult to scrutinize, contest, or assign accountability for errors, biases, or negative unintended consequences.
• Data Governance and Bias: The effectiveness and fairness of AI-GIS applications are heavily dependent on the quality, representativeness, and governance of the input data. Issues of data privacy, security, ownership, and potential biases within datasets (e.g., underrepresentation of certain ecological zones or social groups) can lead to skewed analyses and inequitable outcomes.24
• Deskilling and Neglect of Local Knowledge: Over-reliance on automated AI-GIS systems could lead to the deskilling of local environmental managers and a neglect of valuable traditional ecological knowledge (TEK) and local expertise, which often provide nuanced, long-term understandings of specific socio-ecological contexts.24

The drive to apply AI-GIS to complex environmental challenges can sometimes lead to a form of “technological solutionism,” where intricate socio-ecological and political issues are framed as primarily technical problems amenable to solutions through more data and better algorithms. This risks obscuring the underlying power imbalances, competing interests, and value-based conflicts that are often at the heart of environmental problems.12 While AI-GIS can furnish invaluable data and analytical insights, it cannot supplant the need for robust political negotiation, inclusive stakeholder engagement 103, and addressing the root socio-economic and political drivers of environmental degradation. A critical geographical perspective is essential to ensure that AI-GIS serves as a tool for genuinely democratic and equitable environmental governance, rather than a means of depoliticizing complex issues or reinforcing technocratic control.

Table 6: AI-Enhanced GIS: Applications and Political Implications

4.7 Historical Continuities and Disruptions

Contemporary AI geographies are not ahistorical phenomena; they are deeply interwoven with, and often exacerbate, historical patterns of uneven development and colonial legacies. At the same time, AI’s unique characteristics and rapid diffusion present potential disruptions to these established territorial dynamics, offering both risks of intensified inequality and opportunities for more equitable futures if consciously steered.

Manifestations of Historical Inequalities in AI Geographies: The current global landscape of AI development, deployment, and impact strikingly reflects and reinforces long-standing patterns of inequality established during colonial eras and perpetuated through subsequent phases of uneven global development.5

• “AI Colonialism” or “Robotic Colonisation”: These terms 5 capture how AI technologies, predominantly developed in and controlled by entities in the Global North, extend economic, cultural, and political influence over less-developed nations. This can manifest as technological dependency, erosion of digital sovereignty, and the imposition of external norms and values through AI systems, impacting local labor markets and human rights.30
• Exploitation of Labor: The AI industry relies on a globalized workforce, with tasks like data labeling, annotation, and content moderation often outsourced to low-wage workers in the Global South. These workers frequently face precarious employment conditions, low pay, and significant psychological burdens, mirroring historical colonial labor structures where value was extracted from the periphery to benefit the core.5
• Environmental Burden Shifting: As discussed previously (Sections 4.2 and 4.5), the environmental costs associated with AI—from resource extraction for hardware to the energy demands of data centers and e-waste disposal—are disproportionately borne by communities in the Global South. This pattern of externalizing environmental degradation echoes colonial practices of resource exploitation and the creation of sacrifice zones.5
• Data Colonialism: The extraction of vast amounts of data from populations worldwide, often without equitable compensation or meaningful consent, to train AI models largely controlled by Global North corporations, is another facet of this neocolonial dynamic.13 This data, like raw materials in previous eras, fuels the innovation and economic power of dominant actors.
• Legacy Infrastructure: While direct causal links between colonial-era physical infrastructure (e.g., energy grids, telecommunication lines) and current AI development capabilities in regions like Africa are complex, the broader legacy of colonialism includes underdeveloped infrastructure, educational systems not always aligned with technological advancement, and economic structures that can hinder indigenous AI innovation.31 Current digital divides often map onto these historical infrastructural and developmental disparities.

Unless actively steered towards equitable outcomes, the trajectory of AI development and diffusion tends to follow and amplify these historical pathways of uneven development and colonial power relations. The “AI revolution,” if un critically managed, risks becoming another chapter in the long story of global disparities rather than a genuine force for equitable global transformation.

AI Technology Diffusion and Territorial Development: The diffusion of AI technology itself presents a complex picture. Rogers’s diffusion of innovations theory can be applied to understand AI adoption patterns 109, noting that AI adoption in some sectors, like higher education, is occurring at a significantly faster pace than traditional technological benchmarks, often driven by innovation vendors and social networks rather than top-down institutional strategy.109 The historical development of Generative AI has progressed through distinct stages—from rule-based systems to model-based algorithms, deep generative methodologies, and now large-scale foundation models 3—each with different resource requirements and diffusion characteristics.

Comparing AI’s diffusion with other transformative technologies, like the printing press 7 or the technologies of the Industrial Revolution 4, highlights AI’s potentially greater speed and scale of impact, but also similar tendencies to create new divides. Geographically, AI diffusion shows a dual potential:

• Concentration: The dominant trend is the strengthening of existing high-tech clusters and global AI hubs, where capital, talent, and data create self-reinforcing cycles of innovation.18
• Dispersion: However, AI’s nature as a technology that can, in theory, be applied almost anywhere with potentially lower costs for geographical distribution 110 offers a disruptive potential. Highly automated and remotely managed facilities (“smart factories”) might locate in peripheral “warehousing hubs” if adequate digital and material infrastructure exists.110 This suggests that while AI could disrupt traditional core-periphery dynamics, realizing this potential for equitable development in peripheral regions is contingent on overcoming significant infrastructural, skill, and governance barriers.31 Without proactive investment and strategy, peripheral regions may simply become sites for automated, remotely-controlled operations with limited local benefit, rather than centers of indigenous AI innovation.

Strategies for More Equitable AI Benefit Distribution: Fostering a more equitable distribution of AI’s benefits across diverse territories requires deliberate and multi-pronged strategies:

• Investment in Global South Capabilities: This includes significant investment in digital infrastructure (connectivity, computing power), fostering AI talent and literacy through education and training programs, and supporting the development of localized, culturally relevant AI training data and offline AI applications tailored to specific needs in sectors like health, agriculture, and education.32
• Inclusive AI Development Practices: Actively building representative datasets, meaningfully engaging local communities and intended users as co-creators in the AI design process, prioritizing transparency in AI models and their data sources, and conducting thorough risk and impact assessments are crucial.112
• Strengthening Digital Sovereignty and Governance: Empowering Global South nations to develop and enforce their own data governance frameworks, ensuring data sovereignty, and participating equitably in global AI governance forums is essential.5 This includes establishing local AI Safety Institutes and mandating Human Rights Impact Assessments for AI deployments.111
• Global Cooperation and Technology Transfer: Enforcing global labor protections and environmental standards across AI supply chains, ensuring fair technology transfer agreements, and providing financial and technical support for AI research and infrastructure development in the Global South are vital components of a more equitable global AI ecosystem.5

Addressing these historical continuities and leveraging AI’s disruptive potential for positive change requires a conscious break from past trajectories through targeted policies, international cooperation, and a commitment to spatial and social justice in the age of AI.

5. SWOT Analysis

This SWOT analysis synthesizes the findings from the preceding sections to provide a structured overview of the strengths, weaknesses, opportunities, and threats associated with Artificial Intelligence from a geographical perspective. This lens highlights how AI’s impacts and potentials are spatially differentiated and context-dependent.

Strengths (of AI from a geographical/societal benefit perspective):

• Enhanced Environmental Monitoring and Management: AI, particularly when integrated with GIS (GeoAI), offers powerful tools for monitoring ecosystems, tracking climate change impacts, managing natural resources, and improving disaster response capabilities.10 This can lead to more informed environmental governance.
• Optimization of Urban Systems: AI is being deployed in smart cities to optimize urban functions such as traffic management, public transport, waste collection, and energy distribution, potentially leading to increased efficiency, sustainability, and quality of urban life.42
• Locally-Adapted Solutions for Development Challenges: There is a growing potential for developing AI applications tailored to specific needs in the Global South, particularly in sectors like healthcare (e.g., diagnostics in remote areas), agriculture (e.g., precision farming, disease detection), and education, offering pathways to address persistent development challenges.53
• Increased Efficiency and Resource Savings: Across various sectors, AI can automate tasks, optimize processes, and improve predictive capabilities, which, if managed effectively and equitably, could lead to significant resource savings and productivity gains.77

Weaknesses (of current AI development/deployment from a geographical perspective):

• Significant and Unevenly Distributed Environmental Footprint: AI’s material infrastructure (data centers, hardware) has a substantial environmental impact, including high energy and water consumption, land transformation, greenhouse gas emissions, resource depletion from mining, and e-waste generation. These burdens are often disproportionately borne by vulnerable regions and communities.1
• Geographical Concentration of Development and Power: AI research, investment, and control are highly concentrated in a few global hubs, primarily in the Global North. This creates significant power imbalances, fosters technological dependencies, and limits the ability of peripheral regions to shape AI’s trajectory or benefit equitably.9
• Embedded Biases and Algorithmic Injustice: AI systems, often trained on biased or unrepresentative data, can perpetuate and even amplify existing social and spatial inequalities, leading to discriminatory outcomes in areas like policing, resource allocation, and access to services.4
• Lack of Transparency and Accountability: The “black box” nature of many AI algorithms makes it difficult to understand their decision-making processes, hindering transparency, contestability, and accountability, especially when errors or harms occur.24
• Persistent Digital Divide: Unequal access to digital infrastructure, AI technologies, data, and relevant skills (AI literacy) exacerbates the digital divide, preventing many communities, particularly in the Global South and marginalized areas within the Global North, from fully participating in or benefiting from the AI revolution.4

Opportunities (for geography/society in shaping AI):

• Developing Culturally-Situated and Indigenous AI Paradigms: There is a significant opportunity to move beyond dominant Western-centric AI models by fostering the development of AI systems that are grounded in local cultural contexts, values, and Indigenous epistemologies, leading to more diverse, equitable, and relevant AI solutions.15
• Leveraging AI for Sustainable Development Goals (SDGs): If developed and deployed ethically and inclusively, AI has the potential to accelerate progress towards many SDGs, including those related to health, education, environmental sustainability, and poverty reduction.6
• Fostering Digital Sovereignty and Technological Autonomy: Targeted policies, capacity-building initiatives, and international cooperation can empower nations in the Global South to enhance their digital sovereignty, develop local AI ecosystems, and reduce technological dependencies.31
• Informing Equitable and Sustainable AI Governance: Geographical analysis, with its focus on spatial justice, environmental impacts, and territorial dynamics, can provide critical insights for crafting more effective, equitable, and sustainable AI governance frameworks at local, national, and global scales.1
• Enhancing Participatory Governance: AI tools, if designed appropriately, could facilitate new forms of citizen participation in urban planning, environmental decision-making, and community development, empowering local voices.103

Threats (posed by AI from a geographical perspective):

• Exacerbation of Historical Colonial Patterns and Global Inequalities: Without deliberate intervention, AI risks deepening existing global divides through mechanisms like “AI colonialism,” data extraction, labor exploitation, and the concentration of benefits in the Global North, further marginalizing the Global South.5
• Accelerated Environmental Degradation and Resource Depletion: The unchecked growth of AI’s resource-intensive infrastructure could lead to irreversible environmental damage, exacerbate climate change, and intensify competition for scarce resources like water and critical minerals.1
• Erosion of Privacy and Expansion of Spatial Surveillance: The proliferation of AI-powered surveillance technologies in urban and other spaces poses significant threats to individual privacy, civil liberties, and can lead to discriminatory practices and social control.85
• Uneven Economic Disruption and Job Displacement: Automation driven by AI is likely to cause significant labor market transformations, with potentially severe and geographically uneven job displacement and economic disruption, particularly affecting low-skilled workers and regions heavily reliant on automatable industries.4
• Misuse in Conflict and Security: The development and deployment of AI in military applications, including lethal autonomous weapon systems (LAWS), pose profound threats to global peace and security, with potentially devastating humanitarian consequences, particularly in politically unstable regions.6

A central tension emerges from this analysis: AI simultaneously presents significant threats to geographical equity and environmental sustainability (as highlighted in Weaknesses and Threats) and offers powerful opportunities to address these very same issues (as seen in Strengths and Opportunities). The ultimate outcome is not predetermined but hinges critically on proactive governance, the establishment of robust ethical frameworks, and conscious design choices that prioritize social and environmental well-being.77 Furthermore, the balance of AI’s strengths, weaknesses, opportunities, and threats is not uniform globally but varies significantly across different geographical contexts and scales. An AI application that constitutes an opportunity in a technologically advanced, well-resourced urban hub in the Global North might manifest as a threat or exacerbate weaknesses in a resource-constrained rural community in the Global South.5 This geographical variability underscores the insufficiency of generic, one-size-fits-all assessments of AI and mandates the need for place-based policies and context-sensitive evaluations of its multifaceted impacts.

6. Conclusion and Future Directions

This report has undertaken a comprehensive investigation of Artificial Intelligence through the analytical lens of geography, exploring its intricate spatial, territorial, and environmental dimensions. The analysis reveals AI not as a disembodied technological force, but as a phenomenon deeply embedded within, and actively reshaping, geographical contexts across multiple scales. Key findings underscore the profound geographical unevenness in AI’s development and deployment, with innovation and benefits largely concentrated in a few global hubs, often at the expense of peripheral regions which bear disproportionate environmental and social burdens. The material infrastructure of AI—from resource-intensive data centers to global mineral supply chains—exacts a significant environmental toll, transforming landscapes and raising urgent questions of sustainability and justice. AI is further reconfiguring urban spaces through smart city initiatives and algorithmic governance, creating new efficiencies but also new forms of surveillance and potential for spatial injustice. Critically, the dominant paradigms of AI often reflect Western epistemologies, necessitating the exploration of culturally-situated and Indigenous AI alternatives to foster genuine inclusivity. The extractive geographies underpinning AI hardware echo historical colonial patterns, demanding a reckoning with ongoing environmental injustices. While AI-enhanced Geographic Information Systems (GIS) offer transformative potential for environmental monitoring and governance, their application is fraught with political implications concerning power, equity, and access. Ultimately, contemporary AI geographies demonstrate both striking continuities with historical patterns of uneven development and the potential for disruptive new territorial configurations.

The contribution of geographical analysis to understanding AI is multifaceted and indispensable. By foregrounding space, place, scale, and environment, geography moves beyond purely technical or economistic accounts to reveal AI’s hidden materialities, its embeddedness in social and political power structures, and its place-based specificities. Geographical perspectives are crucial for “grounding” AI, making its abstract processes and global flows tangible in specific locations and communities. This critical spatial lens enriches interdisciplinary conversations by highlighting the distributional consequences of AI, the contestations over territorial control in the digital age, and the urgent need for environmentally sound and socially just AI pathways. Geography’s role must extend beyond critique; it has a vital part to play in proactively shaping more equitable and sustainable AI futures through engagement in co-design processes, policy formulation, and public education.

The rapid evolution of AI and its deepening integration into society open up numerous avenues for future geographical research:

1. Longitudinal Environmental Impact Studies: There is a pressing need for detailed, multi-scalar longitudinal studies that track the cumulative environmental impacts of AI infrastructure (data centers, hardware lifecycle) across diverse ecosystems and geographical contexts. This includes further research into the full “eco-political economy of AI” 12, particularly the social and environmental costs of e-waste from AI hardware.
2. Geographies of Alternative AI: More in-depth qualitative and ethnographic case studies are required to understand the development, implementation, and impact of community-led, culturally-situated, and Indigenous AI initiatives, especially in the Global South. This research should explore how these alternatives challenge dominant paradigms and foster local empowerment.
3. AI and Rural Geographies: While much attention has focused on AI in urban contexts, its implications for rural areas—including agriculture, resource management, access to services, and economic transformation—warrant significantly more geographical investigation.
4. Spatially Just AI Governance: Developing robust geographical frameworks and practical tools for ethical AI governance that explicitly incorporate principles of spatial justice is a critical research frontier. This includes examining how AI can be used to mitigate, rather than exacerbate, spatial inequalities.
5. The Geography of AI Ethics: Future research should critically examine the “geography of AI ethics” itself: how are ethical frameworks for AI developed, by whom, and in which geographical contexts? How do these frameworks travel globally, and how are they adapted, resisted, or reinterpreted in different socio-cultural settings?
6. Participatory GeoAI: Exploring the potential of GeoAI and related AI tools to facilitate more participatory forms of environmental governance, citizen science, and community-based resource management is essential for democratizing access to spatial information and decision-making.

Engaging with the complex challenges and opportunities presented by AI can also serve as a catalyst for reinvigorating geographical theory and practice. Understanding AI’s distributed agency, its capacity to reconfigure space-time, and its role in producing new forms of algorithmic territory may push existing theoretical frameworks in novel directions. The imperative to analyze vast, dynamic, and often real-time geospatial data related to AI’s impacts can spur further innovation in GeoAI methodologies within the discipline. Thus, the study of AI is not merely an application of geographical thought to a new object; it is a reciprocal relationship where AI itself offers a rich empirical and conceptual terrain that can stimulate the evolution of geography as a critical and engaged discipline in the 21st century.

Referências citadas

1. Geography for AI sustainability and sustainability for GeoAI - Taylor & Francis Online, acessado em maio 9, 2025, https://www.tandfonline.com/doi/full/10.1080/15230406.2025.2479796?src=
2. [AM-08-093] Artificial Intelligence Approaches | By ITC, University of Twente - Living Textbook, acessado em maio 9, 2025, https://gistbok-ltb.ucgis.org/27/concept/8009
3. Generative artificial intelligence: a historical perspective - PMC, acessado em maio 9, 2025, https://pmc.ncbi.nlm.nih.gov/articles/PMC11970245/
4. AI And Social Inequality: Technology’s Role In Widening Gaps - RikiGPT, acessado em maio 9, 2025, https://rikigpt.com/ai-and-social-inequality-technologys-role-in-widening-gaps/
5. Tech Imperialism Reloaded: AI, Colonial Legacies, and the Global …, acessado em maio 9, 2025, https://www.e-ir.info/2025/02/17/tech-imperialism-reloaded-ai-colonial-legacies-and-the-global-south/
6. Artificial Intelligence (AI) - the United Nations, acessado em maio 9, 2025, https://www.un.org/en/global-issues/artificial-intelligence
7. Artificial Intelligence - AI | UNESCO, acessado em maio 9, 2025, https://www.unesco.org/en/artificial-intelligence
8. Full article: Regional artificial intelligence and the geography of environmental technologies: does local AI knowledge help regional green-tech specialization? - Taylor & Francis Online, acessado em maio 9, 2025, https://www.tandfonline.com/doi/full/10.1080/00343404.2022.2092610
9. The geography of generative AI’s workforce impacts will likely differ …, acessado em maio 9, 2025, https://www.brookings.edu/articles/the-geography-of-generative-ais-workforce-impacts-will-likely-differ-from-those-of-previous-technologies/
10. Mapping the landscape and roadmap of geospatial artificial …, acessado em maio 9, 2025, https://sdl.gis.harvard.edu/publications/mapping-landscape-and-roadmap-geospatial-artificial-intelligence-geoai
11. GEO:AI - Geography, acessado em maio 9, 2025, https://geog.ufl.edu/geoai/
12. An eco-political economy of AI to understand the complexities of its …, acessado em maio 9, 2025, https://cepr.org/voxeu/columns/eco-political-economy-ai-understand-complexities-its-environmental-costs
13. Decoloniality In AI → Term - Prism → Sustainability Directory, acessado em maio 9, 2025, https://prism.sustainability-directory.com/term/decoloniality-in-ai/
14. Beyond Distribution: Critique of Spatial Justice Theories—Case Study of Shanghai’s 15-Minute City - MDPI, acessado em maio 9, 2025, https://www.mdpi.com/2073-445X/14/1/189
15. 17. Epistemological Clash: AI and Ways of Knowing, acessado em maio 9, 2025, https://digitalcommons.usu.edu/cgi/viewcontent.cgi?article=1024&context=teachingai
16. Full article: Why does urban Artificial Intelligence (AI) matter for …, acessado em maio 9, 2025, https://www.tandfonline.com/doi/full/10.1080/02723638.2024.2329401
17. What is spatial justice? - FutureLearn, acessado em maio 9, 2025, https://www.futurelearn.com/info/courses/african-cities/0/steps/48325
18. The Hidden Multiplier: Unraveling the True Cost of the Global AI …, acessado em maio 9, 2025, https://technologyandsociety.org/the-hidden-multiplier-unraveling-the-true-cost-of-the-global-ai-skills-gap/
19. China’s Strategic Control Over Rare Earths: Global Supply Chain Implications, acessado em maio 9, 2025, https://www.orfonline.org/expert-speak/china-s-strategic-control-over-rare-earths-global-supply-chain-implications
20. a call to (re)think algorithmic harms and resistance through environmental lens The supply chain capitalism of AI - ResearchGate, acessado em maio 9, 2025, https://www.researchgate.net/publication/389421031_The_supply_chain_capitalism_of_AI_a_call_to_rethink_algorithmic_harms_and_resistance_through_environmental_lens_The_supply_chain_capitalism_of_AI_a_call_to_rethink_algorithmic_harms_and_resistance_thr
21. Explained: Generative AI’s environmental impact | MIT News, acessado em maio 9, 2025, https://news.mit.edu/2025/explained-generative-ai-environmental-impact-0117
22. Can AI Save the Amazon? The Politics of AI Environmentalism in Colombia - Leiden University, acessado em maio 9, 2025, https://www.staff.universiteitleiden.nl/events/2025/04/the-politics-of-ai-environmentalism-in-colombia
23. AI and climate resilience governance - PMC, acessado em maio 9, 2025, https://pmc.ncbi.nlm.nih.gov/articles/PMC11112607/
24. Full article: Artificial intelligence and machine learning-powered GIS …, acessado em maio 9, 2025, https://www.tandfonline.com/doi/full/10.1080/19475683.2025.2473596?src=exp-la
25. On actor-network theory and algorithms: ChatGPT and the new …, acessado em maio 9, 2025, https://www.researchgate.net/publication/371944379_On_actor-network_theory_and_algorithms_ChatGPT_and_the_new_power_relationships_in_the_age_of_AI
26. Bruno Latour and Artificial Intelligence | Tecnoscienza – Italian …, acessado em maio 9, 2025, https://tecnoscienza.unibo.it/article/view/18359
27. Critical Data Studies - Edinburgh Futures Institute, acessado em maio 9, 2025, https://efi.ed.ac.uk/ecosystem/critical-data-studies/
28. Critical Data Studies Lab, acessado em maio 9, 2025, https://sites.utexas.edu/data/
29. Artificial Intelligence Colonialism: Environmental Damage, Labor Exploitation, and Human Rights Crises in the Global South - Project MUSE, acessado em maio 9, 2025, https://muse.jhu.edu/article/950958
30. The Global Reach of AI: A Postcolonial Analysis of Technological …, acessado em maio 9, 2025, https://ijsrst.com/index.php/home/article/view/IJSRST24116501
31. How Leaders in the Global South Can Devise AI Regulation That …, acessado em maio 9, 2025, https://institute.global/insights/tech-and-digitalisation/how-leaders-in-the-global-south-can-devise-ai-regulation-that-enables-innovation
32. Manus AI and the Governance Gap: Will the Global South Be Left Behind?, acessado em maio 9, 2025, https://moderndiplomacy.eu/2025/03/14/manus-ai-and-the-governance-gap-will-the-global-south-be-left-behind/
33. An intellectual history of digital colonialism - Oxford Academic, acessado em maio 9, 2025, https://academic.oup.com/joc/advance-article/doi/10.1093/joc/jqaf003/8078024
34. African Digital Colonialism is the New Face of Worker Exploitation …, acessado em maio 9, 2025, https://www.ictworks.org/african-digital-colonialism/
35. Framework for infusing AI with indigenous epistemologies. This …, acessado em maio 9, 2025, https://www.researchgate.net/figure/Framework-for-infusing-AI-with-indigenous-epistemologies-This-diagram-illustrates-the_fig3_384154199
36. Ethical Co-Development of AI Applications with Indigenous …, acessado em maio 9, 2025, https://www.researchgate.net/publication/391152194_Ethical_Co-Development_of_AI_Applications_with_Indigenous_Communities
37. AI and Feminism: Foundations of a Critical Dialogue on Gender …, acessado em maio 9, 2025, https://www.yorku.ca/cifal/ai-feminism/
38. Women Are Avoiding AI. Will Their Careers Suffer? | Working Knowledge - Baker Library, acessado em maio 9, 2025, https://www.library.hbs.edu/working-knowledge/women-are-avoiding-using-artificial-intelligence-can-that-hurt-their-careers
39. Female perspectives on algorithmic bias: implications for AI researchers and practitioners | Emerald Insight, acessado em maio 9, 2025, https://www.emerald.com/insight/content/doi/10.1108/md-04-2024-0884/full/html
40. diva-portal.org, acessado em maio 9, 2025, https://diva-portal.org/smash/get/diva2:1944442/FULLTEXT01.pdf
41. AI Rivalries: Redefining Global Power Dynamics - TRENDS Research & Advisory, acessado em maio 9, 2025, https://trendsresearch.org/insight/ai-rivalries-redefining-global-power-dynamics/
42. Digital Technology and AI for Smart Sustainable Cities in the Global South: A Critical Review of Literature and Case Studies - MDPI, acessado em maio 9, 2025, https://www.mdpi.com/2413-8851/9/3/72
43. Making good - Ada Lovelace Institute, acessado em maio 9, 2025, https://www.adalovelaceinstitute.org/report/ai-public-good/
44. Artificial Intelligence Index Report 2025 - AWS, acessado em maio 9, 2025, https://hai-production.s3.amazonaws.com/files/hai_ai_index_report_2025.pdf
45. oxfordinsights.com, acessado em maio 9, 2025, https://oxfordinsights.com/wp-content/uploads/2024/12/2024-Government-AI-Readiness-Index-2.pdf
46. Harnessing the Power of Artificial Intelligence for Strategic Economic Development, acessado em maio 9, 2025, https://researchfdi.com/ai-economic-development-investment-attraction/
47. View of Global AI Readiness Clusters: Patterns and Disparities in Governance and Technological Infrastructure, acessado em maio 9, 2025, https://gbis.ch/index.php/gbis/article/view/600/501
48. Why AI uses so much energy—and what we can do about it, acessado em maio 9, 2025, https://iee.psu.edu/news/blog/why-ai-uses-so-much-energy-and-what-we-can-do-about-it
49. The Environmental Impact of AI Data Centres: Addressing the Water Use Challenge - Cotie, acessado em maio 9, 2025, https://www.cotie.co.uk/post/the-environmental-impact-of-ai-data-centres-addressing-the-water-use-challenge
50. Exploring the sustainable scaling of AI dilemma: A projective study of corporations’ AI environmental impacts - arXiv, acessado em maio 9, 2025, https://arxiv.org/pdf/2501.14334?
51. How can nations overcome the great AI divide? | World Economic …, acessado em maio 9, 2025, https://www.weforum.org/stories/2025/01/how-can-all-nations-realize-the-full-potential-of-ai/
52. (PDF) Digital Technology and AI for Smart Sustainable Cities in the …, acessado em maio 9, 2025, https://www.researchgate.net/publication/389619456_Digital_Technology_and_AI_for_Smart_Sustainable_Cities_in_the_Global_South_A_Critical_Review_of_Literature_and_Case_Studies
53. Responsible AI in Global Health: Solutions from the Global South …, acessado em maio 9, 2025, https://publichealth.jhu.edu/center-for-global-digital-health-innovation/responsible-ai-in-global-health-solutions-from-the-global-south
54. USING GEOGRAPHICAL INFORMATION SYSTEMS (GIS) FOR CLIMATE CHANGE REDUCTION AND ADAPTATION: A COMPREHENSIVE REVIEW, acessado em maio 9, 2025, https://ijees.net/images/pdf/HaticeCananGUNGOR1AbdinasirOsmanMOHAMED2USINGGEOGRAPHICALINFORMATIONSYSTEMS(GIS)FORCLIMATECHANGEREDUCTIONANDADAPTATION:ACOMPREHENSIVEREVIEWpage71-80;-b37470a5e8.pdf
55. Artificial intelligence-enabled assessment of urban growth impacts …, acessado em maio 9, 2025, https://www.tandfonline.com/doi/full/10.1080/10225706.2025.2481573?src=exp-la
56. Full article: AI From the South: artificial intelligence in Latin America through the sociotechnical imaginaries of Brazilian tech workers - Taylor & Francis Online, acessado em maio 9, 2025, https://www.tandfonline.com/doi/full/10.1080/14747731.2025.2465166?src=exp-la
57. The Case for ‘Thick Evaluations’ of Cultural Representation in AI - arXiv, acessado em maio 9, 2025, https://arxiv.org/html/2503.19075v1
58. Mapping the Knowledge Base on Sustainable AI: A Systematic Review, acessado em maio 9, 2025, https://www.jriiejournal.com/wp-content/uploads/2025/04/JRIIE-9-2-011.pdf
59. Leveraging artificial intelligence in disaster management: A comprehensive bibliometric review | Wibowo | Jàmbá, acessado em maio 9, 2025, https://jamba.org.za/index.php/jamba/article/view/1776/3377
60. The untold story of the environmental impact of AI and big data - University of York, acessado em maio 9, 2025, https://www.york.ac.uk/news-and-events/news/2025/research/decarbonising-digital-infrastructure/
61. A case study of China’s contact tracing systems in the COVID-19 - University of Washington, acessado em maio 9, 2025, https://geography.washington.edu/sites/geography/files/documents/research/1-s2.0-s0016718522002184-main.pdf
62. From Global Governance to Nationalism: The Future of AI, acessado em maio 9, 2025, https://www.giga-hamburg.de/en/publications/giga-focus/from-global-governance-to-nationalism-the-future-of-ai
63. AI- Driven Territorial Intelligence - IGI Global, acessado em maio 9, 2025, https://www.igi-global.com/viewtitle.aspx?TitleId=354608&isxn=9798369385579
64. (PDF) AI Sovereignty, and Strategy: Navigating the Geopolitical Landscape for African States - ResearchGate, acessado em maio 9, 2025, https://www.researchgate.net/publication/390523034_AI_Sovereignty_and_Strategy_Navigating_the_Geopolitical_Landscape_for_African_States
65. Beyond Big Tech Geopolitics | Transnational Institute, acessado em maio 9, 2025, https://www.tni.org/en/article/beyond-big-tech-geopolitics
66. Spatial Inequality and Development | Oxford Academic, acessado em maio 9, 2025, https://academic.oup.com/book/36181
67. Technology and Innovation Report 2025 - Chapter V: Global collaboration for inclusive and equitable AI - UNCTAD, acessado em maio 9, 2025, https://unctad.org/system/files/official-document/tir2025ch5_en.pdf
68. Submarine Cable Sensing for Strategic Infrastructure Defense and Arctic Deployment., acessado em maio 9, 2025, https://techneconomyblog.com/2025/04/01/submarine-cable-sensing-for-strategic-infrastructure-defense-and-arctic-deployment/
69. Is fibre optic sustainable? - Greenly, acessado em maio 9, 2025, https://greenly.earth/en-gb/blog/industries/is-fibre-optic-sustainable
70. Remote Sensing and Geospatial Analysis in the Big Data Era: A Survey - MDPI, acessado em maio 9, 2025, https://www.mdpi.com/2072-4292/17/3/550
71. Integrating Remote Sensing and Geospatial Big Data for Land Cover and Land Use Mapping and Monitoring - ResearchGate, acessado em maio 9, 2025, https://www.researchgate.net/publication/381143313_Integrating_Remote_Sensing_and_Geospatial_Big_Data_for_Land_Cover_and_Land_Use_Mapping_and_Monitoring
72. Harnessing artificial intelligence (AI) towards the landscape of big earth data: Methods, challenges, opportunities, future directions - ResearchGate, acessado em maio 9, 2025, https://www.researchgate.net/publication/388223443_Harnessing_artificial_intelligence_AI_towards_the_landscape_of_big_earth_data_Methods_challenges_opportunities_future_directions
73. Keeping connected: submarine communications cables and ocean life - UNEP-WCMC, acessado em maio 9, 2025, https://www.unep-wcmc.org/en/news/keeping-connected-submarine-communications-cables-and-ocean-life
74. Water Risk for Data Centers - Waterplan, acessado em maio 9, 2025, https://www.waterplan.com/blog/water-risk-for-data-centers
75. Data Ecofeminism - arXiv, acessado em maio 9, 2025, https://arxiv.org/html/2502.11086v1
76. (PDF) The Environmental Impacts of AI and Digital Technologies - ResearchGate, acessado em maio 9, 2025, https://www.researchgate.net/publication/379041862_The_Environmental_Impacts_of_AI_and_Digital_Technologies
77. Mitigating AI’s environmental impact: A path to … - Tech For Good, acessado em maio 9, 2025, https://www.techforgood.net/thoughtleadership/mitigating-ais-environmental-impact-a-path-to-sustainable-innovation
78. Within Bounds: Limiting AI’s environmental impact - Green Screen Coalition, acessado em maio 9, 2025, https://greenscreen.network/en/blog/within-bounds-limiting-ai-environmental-impact/
79. AI Governance In Sustainability → Term, acessado em maio 9, 2025, https://prism.sustainability-directory.com/term/ai-governance-in-sustainability/
80. Applied Sciences | Special Issue : AI-Based Spatial Planning and …, acessado em maio 9, 2025, https://www.mdpi.com/journal/applsci/special_issues/367K84O0O3
81. Geospatial Intelligence & CV: Applications in Urban Planning - viso.ai, acessado em maio 9, 2025, https://viso.ai/applications/geospatial-intelligence/
82. The Emergence of GeoAI in Planning - Industry Blogs - Esri, acessado em maio 9, 2025, https://www.esri.com/en-us/industries/blog/articles/the-emergence-of-geoai-in-planning
83. Building Urban Resilience Through Smart City Planning: A Systematic Literature Review, acessado em maio 9, 2025, https://www.mdpi.com/2624-6511/8/1/22
84. (PDF) Algorhythmic governance: Regulating the ‘heartbeat’ of a city using the Internet of Things - ResearchGate, acessado em maio 9, 2025, https://www.researchgate.net/publication/321136819_Algorhythmic_governance_Regulating_the_‘heartbeat’_of_a_city_using_the_Internet_of_Things
85. How Governments are Using AI: 8 Real-World Case Studies, acessado em maio 9, 2025, https://blog.govnet.co.uk/technology/ai-in-government-case-studies
86. Ethical Implications of AI in Urban Resource Allocation - Prism → Sustainability Directory, acessado em maio 9, 2025, https://prism.sustainability-directory.com/scenario/ethical-implications-of-ai-in-urban-resource-allocation/
87. Legal and Regulatory Considerations for States Related to Artificial Intelligence, acessado em maio 9, 2025, https://www.nga.org/updates/legal-and-regulatory-considerations-for-states-related-to-artificial-intelligence/
88. Practice and Application of Artificial Intelligence in Urban Decision-Making - MDPI, acessado em maio 9, 2025, https://www.mdpi.com/journal/buildings/special_issues/UZMN9BF5U7
89. AI in Architecture and Urban Design and Planning: Case studies on three AI applications, acessado em maio 9, 2025, https://www.researchgate.net/publication/386348989_AI_in_Architecture_and_Urban_Design_and_Planning_Case_studies_on_three_AI_applications
90. Can AI Solve Food and Energy Crises in the Global South …, acessado em maio 9, 2025, https://moderndiplomacy.eu/2025/03/01/can-ai-solve-food-and-energy-crises-in-the-global-south/
91. conference2023.eaamo.org, acessado em maio 9, 2025, https://conference2023.eaamo.org/papers/EAAMO23_paper_154.pdf?ref=internet.exchangepoint.tech
92. What Ethical Considerations Arise With Battery Production? - Sustainability Directory, acessado em maio 9, 2025, https://sustainability-directory.com/question/what-ethical-considerations-arise-with-battery-production/
93. The Environmental Impacts of Cobalt Mining in Congo | Earth.Org, acessado em maio 9, 2025, https://earth.org/cobalt-mining-in-congo/
94. Supply Chain Mapping: A Key to Sustainability and Transparency …, acessado em maio 9, 2025, https://www.amfori.org/en/news/supply-chain-mapping-a-key-to-sustainability-and-transparency
95. AI Tools Driving Sustainability in Procurement & Supply Chains | GEP Blog, acessado em maio 9, 2025, https://www.gep.com/blog/strategy/ai-role-in-sustainable-supply-chains-decarbonization
96. The hidden cost of AI: Unpacking its energy and water footprint - OECD.AI, acessado em maio 9, 2025, https://oecd.ai/en/wonk/the-hidden-cost-of-ai-energy-and-water-footprint
97. AI-Driven Environmental Justice → Term - Prism → Sustainability Directory, acessado em maio 9, 2025, https://prism.sustainability-directory.com/term/ai-driven-environmental-justice/
98. Geospatial AI - Artificial Intelligence in GIS - Esri, acessado em maio 9, 2025, https://www.esri.com/en-us/geospatial-artificial-intelligence/overview
99. Full article: Artificial intelligence and machine learning-powered GIS for proactive disaster resilience in a changing climate - Taylor & Francis Online: Peer-reviewed Journals, acessado em maio 9, 2025, https://www.tandfonline.com/doi/full/10.1080/19475683.2025.2473596?src=
100. (PDF) The Role of GIS in Disaster Response: Improving Efficiency and Overcoming Challenges - ResearchGate, acessado em maio 9, 2025, https://www.researchgate.net/publication/389524130_The_Role_of_GIS_in_Disaster_Response_Improving_Efficiency_and_Overcoming_Challenges
101. Artificial intelligence in environmental monitoring: in-depth analysis, acessado em maio 9, 2025, https://d-nb.info/1355324793/34
102. (PDF) The Role of GIS in the Study of Sustainable Development and Environmental Management - ResearchGate, acessado em maio 9, 2025, https://www.researchgate.net/publication/387054544_The_Role_of_GIS_in_the_Study_of_Sustainable_Development_and_Environmental_Management
103. GIS in Natural Resource Management – Mapping & Sustainability, acessado em maio 9, 2025, https://gisnavigator.co.uk/gis-in-natural-resource-management/
104. Artificial Intelligence in Aquatic Biodiversity Research: A PRISMA-Based Systematic Review, acessado em maio 9, 2025, https://www.mdpi.com/2079-7737/14/5/520
105. Artificial Intelligence for Sustainable Urban Biodiversity: A Framework for Monitoring and Conservation - arXiv, acessado em maio 9, 2025, https://www.arxiv.org/pdf/2501.14766
106. GIS for Water Resources and Watershed Management (Uses), acessado em maio 9, 2025, https://gisnavigator.co.uk/gis-for-water-resources-and-watershed-management/
107. AI Policy Analysis 2025 | Public Policy with Intelligent Systems - Rapid Innovation, acessado em maio 9, 2025, https://www.rapidinnovation.io/post/ai-agent-public-policy-impact-analyzer
108. The Future of GIS Work in the Age of AI: Evolution, Transformation, and Potential Displacement - Geospatial Training Services, acessado em maio 9, 2025, https://geospatialtraining.com/the-future-of-gis-work-in-the-age-of-ai-evolution-transformation-and-potential-displacement/
109. (PDF) Artificial Intelligence: Diffusion of Innovation and Generational Considerations Higher Education Institutions - ResearchGate, acessado em maio 9, 2025, https://www.researchgate.net/publication/389901448_Artificial_Intelligence_Diffusion_of_Innovation_and_Generational_Considerations_Higher_Education_Institutions
110. (PDF) Reviewing research on regional development in the AI era: new there(s), new actors, and an old call “from cluster to process” - ResearchGate, acessado em maio 9, 2025, https://www.researchgate.net/publication/390414566_Reviewing_research_on_regional_development_in_the_AI_era_new_theres_new_actors_and_an_old_call_from_cluster_to_process
111. (PDF) Bridging the International AI Governance Divide: Key …, acessado em maio 9, 2025, https://www.researchgate.net/publication/391396519_Bridging_the_International_AI_Governance_Divide_Key_Strategies_for_Including_the_Global_South
112. How to Make AI Equitable in the Global South - Stanford Social Innovation Review, acessado em maio 9, 2025, https://ssir.org/articles/entry/equitable-ai-in-the-global-south