Introduction

Geography, as a discipline, is traditionally divided into two main branches: human geography and physical geography. Human geography focuses on spatial phenomena triggered by human activities (Norton and Mercier 2016), while physical geography emphasizes the ecological evolution of space (Pournelle 2013). Between these two paradigms, human geography has been relatively contentious throughout the history of geography, primarily due to debates surrounding its particular academic and scientific nature, which has attracted skepticism and challenges from other fields within geography itself, as well as external disciplines (James 1949). From a spatial perspective, elements such as culture (Fouberg et al. 2015), ethnicity (Anderson 2014), and language (Chapman 2023) exhibit both significant diversity and specificity (Bingham 1999), making it difficult for human geography to establish robust, universally applicable theories (Ley 2014). At present, the only widely recognized theoretical principle in the field is the First Law of Geography (Tobler 2004), yet its applicability to the microscale of spatial analysis remains a subject of debate (Sui 2004). Regarding its distinctiveness, much of the research in human geography on communities, race, and cultural ideologies draws heavily on disciplines such as sociology (Gregory 1994), philosophy (Graham 2013), and political science (Smith 1994). This reliance often makes it difficult to distinguish human geography’s perspective from those of these adjacent disciplines, thereby challenging its perceived influence and recognition among other fields and in universities (Philo 2017).

The cancellation of human geography at Harvard University, on the grounds of lacking scientific rigor (Smith 1987) triggered a trend among other leading universities with similar traditions to abolish the discipline, thereby undermining the academic development of human geography and halting progress in both talent cultivation and specialized research. Second, while spatial research forms the foundation of geography (Geus 2003), human geography has often struggled to delineate its boundaries from adjacent disciplines such as planning (Madanipour and Hull 2017), environmental studies (Koons et al. 1999), and architecture (Ching 2023). This has hindered its ability to demonstrate how spatial perspectives can address human-centric issues effectively (Merriman et al. 2012). Furthermore, the integration of human and geographical spatial relationships continues to perplex scholars in the field. Third, with the advancement of geographic information systems (GIS), human geographers rely increasingly on quantitative methods to analyze phenomena such as population mobility, spatial distribution, and network relations. Traditional qualitative approaches in human geography, such as studies of human behavior and human–environment interactions, have been largely displaced by information technologies. Many scholars now contend that nondata-driven or noninformation-based human geography research has lost both its persuasiveness and practical relevance.

To strengthen the scientific status and influence of human geography, the discipline has increasingly adopted concepts and technical methods from the natural sciences and information technology fields. Over the past 30 years, a research paradigm dominated by quantitative approaches has emerged, with human geographers employing various relational models to uncover spatial phenomena and their underlying causes. Examples include the delineation of urban functional zones, the formation of urban agglomerations, the interplay between urban and rural areas, and the structural dynamics of community networks. However, a review of the past decade reveals a paucity of major societal issues resolved or useful conclusions drawn using the unique analytical methods of human geography. This can be attributed to the following three reasons: 1. Some human geography problems or phenomena identified directly through data-driven approaches lack any tangible impacts on the daily lives of residents in practice. 2. With interdisciplinary research technologies dominating geographical studies, geographers can bypass traditional human geography methods. Geographic research relies increasingly on geoscience data and computational models to produce conclusions, reducing the significance of the research region to merely providing a physical location, and in some cases, scholars select research sites without considering their unique or symbolic significance. 3. The exploration of academic terminology used within human geography often revolves around debates over interpretative authority, while many established definitions and standardized terms originate from other fields or external institutions. For instance, the concept of urban resilience, introduced in 1973 (Meerow et al. 2016), was devised as an academic framework to enhance sustainable urban development and protect urban residents in the face of disasters. Despite 50 years of research, geography has struggled to establish a comprehensive system of definitions, standards, and implementation strategies to support urban development effectively. Although interdisciplinary knowledge has, in recent years, enabled cities to mitigate losses from natural disasters, the presence and influence of human geography within such efforts have diminished progressively.

The fundamental objective of interdisciplinarity lies in leveraging knowledge from other disciplines to enhance and refine one’s own disciplinary knowledge system. In the context of human geography, a pivotal contemporary issue concerns how the discipline can integrate interdisciplinary insights while preserving its distinctive attributes, especially amid the transformative impact of the AI era. The advent of AI is reshaping human geography at an unprecedented pace. Publicly accessible AI technologies such as ChatGPT are increasingly becoming integral components of the research process. Unlike the technological tools of the information age, AI technologies transcend mere facilitation of data retrieval or processing; they increasingly employ algorithmic learning to supplant human and disciplinary modes of cognition and learning. As AI systems progressively assimilate the accumulated knowledge of past scholarship, they demonstrate superior knowledge architectures and research capabilities in numerous domains relative to individual researchers.

It is likely caused to AI’s evolution as a direct extension and iteration of prior information technologies, human geography has predominantly regarded AI as an ancillary tool aimed at augmenting extant capacities such as data processing, image analysis, and text mining, rather than critically engaging with it as a potentially transformative, substitutional knowledge system. Of particular concern is the widespread tendency in research practice to treat AI as an “externalized” technology (Skulmowski 2023), seemingly divorced from the core academic values of human geography such as interpretivism, criticality, and contextualization. Yet, this instrumentalist perspective obscures a crucial reality: AI fundamentally possesses cognitive and evaluative functions capable of partially substituting human geographers within specific contexts. From automated geomorphological labeling through image recognition to sentiment analysis and quantitative representation of socio-spatial atmospheres, AI technologies are progressively permeating and reshaping the epistemic foundations upon which human geography constructs notions of “place meaning” and “spatial experience” (Torrens 2018). In the AI era, human geography must not only harness AI technologies to redress previous research limitations but also develop more integrative methodologies that preserve the irreplaceable scholarly value inherent to the discipline itself.

Literature review

Although human geography has faced considerable controversy in its modern development, it has consistently played a positive and significant role in the historical progress of humanity. The term “human geography” originated in the 19th century, evolving from Anthropogeography, a concept introduced by German scholar Friedrich Ratzel (Bassin 1987). As humanity expanded its global understanding through the Age of Exploration and colonialism, human geography became a vital channel for understanding the places, cultures, and characteristics of nations worldwide. In 1920, Carl Sauer introduced the concept of the cultural landscape to human geography, establishing its core focus: the interaction between human societies and their environments, which shapes cultural landscapes (Fleming 1984). Beginning in the 1950s, the Quantitative Revolution (Barnes 1994; Burton 1963) significantly influenced the field of human geography, introducing technological methods that diversified its research approaches. This diversification required interdisciplinary support, and by the 1980s, the emergence of Critical Geography began challenging traditional human geography methods, such as surveys, short-term field studies, and interviews (Davies et al. 2014). The widespread adoption of GIS marked a turning point, signaling that the interdisciplinary knowledge informing human geography would no longer be limited to social sciences. Knowledge systems and research methods from the natural and engineering sciences have increasingly been integrated into human geography, thereby posing challenges to its traditional research paradigms. The advent of the AI era has intensified these challenges, as AI technologies gradually replace many traditional approaches to studying societal issues. This trend presents a paradox: while integrating interdisciplinary knowledge, human geography has not gained strength through this diversification, but instead faces the growing risk of being subsumed or even replaced by other disciplines.

Human geography today employs both qualitative and quantitative analytical approaches. Qualitative analysis is a research method based on observation and investigation, aimed at understanding and interpreting social phenomena, human behavior, and cultural meanings (Gibbs 2014). It typically involves the collection and analysis of non-numerical data such as text, images, interviews, and observations, focusing on exploratory, descriptive, and interpretive studies that emphasize a deep understanding of complex phenomena. However, due to its high degree of individual subjectivity and the small sample sizes often involved, the findings of qualitative analysis are somewhat challenging to generalize (Cope 2010). As a result, some human geographers have adopted quantitative analysis methods to investigate and explore the causes of various social phenomena (Brunsdon 2016). Quantitative analysis is a numerical data-driven approach that aims to describe, explain, and predict spatial phenomena and human behavior using statistical and mathematical models. It involves data collection, measurement, computation, and statistical analysis to identify relationships, trends, and patterns in data, make generalizable inferences, and test hypotheses. Quantitative analysis emphasizes causality, regularity, and predictive research, focusing on the objectivity of data and related statistical significance. Common methods include statistical analysis (e.g., regression analysis, analysis of variance), spatial analysis (e.g., GIS, spatial autocorrelation analysis), and model construction (e.g., gravity models, network analysis) (Miller and Wentz 2003). However, research dominated by quantitative methods often neglects social contexts and subjective meanings. More critically, some scholars, when faced with data limitations, rely heavily on assumptions and predetermined conclusions to conduct their studies (Johnston et al. 2019).

To better integrate qualitative and quantitative research methods, human geographers often combine various dimensions and scales in their studies, striving to ensure that the research process and conclusions are more objective and thorough (Elwood 2010). In human geography, research dimensions typically encompass social, economic, and environmental aspects, enabling geographers to develop a holistic and in-depth understanding of the complex interactions between human activities and geographic spaces (Ikwuyatum 2018; Hägerstrand 2019). This use of dimensions helps uncover the integrated characteristics of geographic phenomena, providing a rich perspective for theoretical development and practical application in geography. Within this framework, scholars employ different scales or multiscalar analyses to explore the interactions between human activities and geographic spaces. These analyses often consider both spatial and temporal scales (Campbell 2018), ranging from micro to macro levels, such as individuals, communities, localities, regions, nations, and even up to the global scale. Such an approach reveals the geographic features and processes operating across different levels and dimensions (Fouberg et al. 2015). The flexibility of research scales in human geography allows scholars to select the most suitable scale of analysis, from personal to global, depending on the research topic and data characteristics. Each scale offers a unique perspective that uncovers specific geographic processes and phenomena, thereby supporting a deeper understanding of the research subject. Furthermore, multiscalar analysis provides a comprehensive and multidimensional perspective, uncovering the intricate geographic processes and interactions between human activities and diverse spatial environments. This approach enables geographers to capture dynamics and variations from global to local scales and from long-term historical changes to short-term developments. By integrating multiple scales, scholars can better understand the interconnectedness and diversity of human–geographic phenomena (Marston et al. 2017). Recognizing that the processes observed at one scale may differ from, or even contradict those observed at another, the multiscalar approach not only enriches the theoretical framework of human geography but also enhances the robustness and practical applicability of its findings, aiding the resolution of complex real-world problems (Buzar 2008). However, while the importance of integrating multiple dimensions and scales has been acknowledged within human geography, the discipline’s lack of a distinct and cohesive research system has led to a diminished sense of its unique identity and research value during the study process (Dymitrow and Brauer 2018).

As AI technologies begin to exert a profoundly disruptive influence on the field of human geography, the discipline has increasingly recognized and engaged with these transformative impacts (Janowicz et al. 2022). Cutting-edge research topics such as the interactive dynamics among AI, humans, and space—often conceptualized as the “AI–Human–Space” (Lai et al. 2023) nexus—have become prominent in scholarly publications and academic conferences. Among these developments, the concept of the virtual scenario stands out as a key hallmark of AI’s influence on human geography in the contemporary era. A “virtual scene” refers to an immaterial spatial representation constructed through digital technologies, which synthesizes perceptual imagery, interactive logics, narrative structures, and data modeling (Weigle et al. 2024). This integrative process serves to simulate or generate an individual’s scene-based experience situated within a specific spatiotemporal context.

The preliminary application of virtual scene within human geography can be traced back to the early information age, primarily manifested through the simulation and reproduction of remote spaces via internet-based platforms, such as virtual meeting rooms and online exhibition halls (Dixon 2006). During this phase, virtual scene partially transcended the geographical constraints of real–world scene, enabling individuals situated at different times and locations to share a common digital environment and thereby facilitate functions such as communication, display, and education. However, this early virtual scene remained largely confined to functional “place substitutions,” lacking mechanisms for adaptive evolution and behavioral nesting, and thus exhibiting limited capacity for social impact and spatial restructuring (Lange 2001).

With the advent of the AI era, virtual scenarios have experienced a qualitative leap. Leveraging advancements in artificial intelligence—particularly large-scale models, generative algorithms, and multimodal perception systems—virtual scenarios have evolved beyond mere image-based information interfaces to become generative spatial structures (Yang and Zhen 2024). Individuals can now access diverse services at a single physical location, while fluidly transitioning between virtual and physical realities and engaging in multifaceted social activities across multiple scenes. This highly malleable interactive logic positions virtual scene as a “spatial behavior medium,” thereby redefining the interactive relations among humans, objects, and places (Shen et al. 2018).

In summary, the integration of interdisciplinary knowledge has transformed human geography into a complex and diverse academic field, encompassing multiple interconnected dimensions, scales, and diverse research methods. The AI era now presents both unprecedented opportunities and significant challenges for human geography (Gao 2024). Developing a well-suited research framework for the AI era not only enables human geographers to conduct and complete their research but also facilitates their exploration and interpretation of complex social phenomena and the spatial distribution and interactions of human activities from a human–geographic perspective. A thoughtfully designed research framework allows human geographers to approach complex geographic phenomena from multiple angles and levels, advancing their studies beyond mere descriptive analyses to the realms of theoretical construction and practical guidance. This paper proposes the Space–Scene–Scenario (3S) framework, designed to balance the preservation of human geography’s disciplinary uniqueness with the current demands of interdisciplinary integration. It will detail the process of constructing the Space–Scene–Scenario (3S) research framework, providing scholars with a systematic and scientific perspective on the research process in human geography. Furthermore, this study aims to highlight the discipline’s unique contributions to both theoretical inquiry and practical applications while also supporting interdisciplinary endeavors in related fields.

The Space–Scene–Scenario (3S) in human geography

Definitions of Space, Scene, and Scenario (3S)

“Space” is a core concept in human geography (Cox 1995), encompassing not only its physical existence but also its social, economic, and cultural dimensions. The study of space in human geography is highly diverse, ranging from analyses of spatial forms and characteristics to investigations of operational mechanisms, such as the formation of urban spatial structures, processes of spatial agglomeration, and patterns of spatial connectivity.

In other disciplines, space is merely a form or context; however, in human geography, it serves as the foundation of the discipline. The challenge in human geography lies in transforming “space” into “spatiality” (Tversky et al. 1999). In contemporary human geography, assigning “spatiality” to “space” represents a critical academic challenge. While space functions as the backdrop and medium for geographical phenomena, spatiality describes the relationship between these phenomena and the research objectives, thereby imbuing space with distinct attributes. With the rise of interdisciplinary approaches, the demand for spatial research in human geography has diminished. Earlier spatial studies often limited their scope to identifying the research area’s location and the elements or phenomena associated with that location (Shafer and Eisenach 2000). However, spatial phenomena initiated by individuals in space are inherently subjective, requiring human geography to analyze these phenomena through a deep understanding of the social, cultural, economic, and political dimensions embedded within space. In contemporary human geography, the process of assigning “spatiality” to “space” requires comprehensive consideration across multiple dimensions, including theoretical construction, multiscale analysis, spatial representation, empirical research, and policy practice. Spatiality is shaped through the material and social production of space. Henri Lefebvre’s theory of the production of space (Schmid 2008) paved the way for examining how space is constructed by social forces and how these forces operate and are reproduced within spatial contexts. Moreover, human geography must explore the interactions across local, regional, and global scales, addressing the impact of globalization on local economies and cultures and the influence of localized social movements on national policies.

After endowing space with spatiality, it is necessary to conduct a more detailed analysis and assign thematic significance to the relationship between humans and space. Previous studies on space can generally be categorized into two main approaches: spatial structure analysis and functional zoning, where space is segmented into different functional regions for specific research purposes. However, the advent of AI technology has fundamentally transformed both the demands and methodologies of spatial research in human geography. First, AI has altered the traditional assumption that spatial development, as exemplified by urban agglomerations, necessitates physical concentration. The increasing prominence of polycentric and decentralized spatial structures challenges conventional theories of agglomeration economies, thereby reshaping the trajectory of regional socio-economic development. Second, in the AI era, there is a growing demand for multi-functional, integrated spaces within individual regions. This shift poses significant challenges to traditional functional or administrative zoning approaches, which often struggle to accurately delineate spatial boundaries and fail to comprehensively capture complex spatial phenomena. To better analyze human activities within space, the Scene framework provides a more effective alternative to functional zoning. Unlike functional zoning, which rigidly classifies space based on predefined uses, Scenes retain spatial complexity and integration while simultaneously incorporating a thematic dimension, making them a more adaptable analytical unit for understanding spatial dynamics in the AI era.

The concept of “scene” has been widely applied across various disciplines, each with its own interpretations and applications. In urban planning, a scene often refers to different spatial development models or planning schemes used to simulate potential future trajectories of urban or regional development (Ma et al. 2021). In literature and the arts, scenes are typically employed to describe stage settings and construct narratives (Tako et al. 2019). With the advent of the AI era, the application of virtual scenes has proliferated, with fields including computer science and artificial intelligence utilizing virtual interaction and virtual reality technologies to enable users to experience multiple distinct scenes within the same location. In recent years, as AI technologies have advanced, disciplines such as medicine have increasingly incorporated virtual reality to generate diverse scenes for training and problem-solving (Ulbrich et al. 2015). In human geography, scene primarily concerns the static perspective of physical space, virtual environments, environmental configurations, and the built environment. It serves as a critical nexus linking spatiality with research themes (Zhang et al. 2018). As an extension of spatial analysis, the concept of scene enables researchers to segment space across multiple scales and assemble heterogeneous elements in alignment with specific research objectives. In some cases, the entire geographical space may constitute a scene; in others, a scene may refer only to a localized setting relevant to the research theme. For instance, in examining the diversification of residents in large metropolitan communities, the metropolis functions as the overarching research space, while a researcher may designate a single neighborhood as the scene. Alternatively, depending on the nature of the research and data requirements, the scene could encompass multiple selected communities or even all neighborhoods within the city.

Unlike conventional spatial delineations such as functional zones or administrative regions, the segmentation of scenes is more human-centered and flexible, particularly for thematic areas such as commercial districts, metropolitan areas, or cultural zones, which otherwise may lack clearly defined boundaries. Moreover, scene-based analysis enables a comprehensive representation of the core themes of the research target. As cities and communities increasingly demand multifunctionality and integration, a focus solely on single-function zones is insufficient to encapsulate the complexity of spatial characteristics and to conduct holistic research. During the thematic analysis of scenes, the role of the scene in human geography emerges as a pivotal element. Scenes not only function as spatial divisions based on varying themes but also respond to humanity’s deeper expectations for integrated and thematic spaces, uncovering the complex interplay between spatial functions and societal significance. In the AI era, the role of virtual scenes within spatial studies has grown increasingly prominent. Virtual scenes enable individuals to access diverse services within the same physical location while simultaneously participating in distinct activities across both real and virtual environments, creating a new medium for human–space interactions (Fan et al. 2024). Against this backdrop, utilizing the scene as a bridge between humans and space in the AI era deepens the thematic dimensions of spatial research while also enriching the multidimensional representation of human factors, thereby highlighting the close connection between people and their environments. Additionally, AI technologies imbue scenes with cross-regional characteristics. Virtual scenes can connect geographically nonadjacent areas that share similar themes, expanding their influence, fostering interaction between scenes, and promoting spatial integration. This dynamic interaction forms a network of thematic scenes across diverse spaces, redefining traditional geographic boundaries significantly. In specific research practices, scholars typically identify a clear research theme and focus on particular populations according to their objectives. As research goals become more refined, sample sizes tend to decrease. Even within the same spatial context, individuals often experience vastly different scenarios in the same scene due to variations in cultural background, gender, or income. These individualized differences not only enrich the connotations of scene studies but also provide diverse perspectives for understanding the complex relationship between humans and space.

A scenario typically refers to an imagined or hypothetical situation used to discuss potential developments, plan actions, or forecast future conditions (Spaniol and Rowland 2019). Compared to context, which focuses on explaining why a particular event occurs or how a concept should be understood, scenario places greater emphasis on simulating or projecting events under different conditions. In human geography, the concept of scenario emphasizes the dynamic interactions between specific population groups and the scenes in which they are situated. It represents a concrete manifestation of how behaviors emerge through the interplay between human agency and spatial configurations under a given research theme. Scenario-based research not only involves the deliberate construction of thematic spatial settings to observe human behavior and its resulting social and spatial phenomena but also aims to simulate or anticipate potential developments and risks associated with such scenes.

For example, in studying the adaptive behaviors of vulnerable urban populations under extreme climate events, researchers may define “prolonged summer heatwaves” as a scenario variable and focus on target groups such as the elderly or low-income residents. By simulating responses across different communities—such as the opening of cooling centers, increasing green space coverage, or investing in thermal infrastructure—researchers can analyze spatial mobility patterns, risk-avoidance strategies, and the mobilization of social networks within this scenario. In this context, the scenario is not merely a predictive model of potential heat-related risks, but an experimental construct to examine the behavior–environment interaction processes among older adults. It reveals how factors such as policy interventions, social capital, and spatial accessibility reshape behavioral choices and spatial arrangements, producing diverse socio-geographical outcomes.

In traditional human geography research, scenario-based analyses have predominantly focused on collective population dynamics, such as demographic distribution, urbanization processes, and regional development. This has often led to conceptual ambiguity and conflation between scenario and scene in the research process. A key reason for this confusion lies in the limitations of earlier information technologies, which primarily served specific demographic groups and lacked the capability to differentiate individuals effectively. However, with the rapid advancement of AI technologies, these constraints are being progressively overcome. For instance, AI-driven tools such as ChatGPT can now generate customized responses based on users’ backgrounds, significantly enhancing the accuracy and personalization of individual-level analysis (Yang and Zhen 2024). This technological progress effectively resolves the long-standing tension in traditional research between large-scale statistical approaches and qualitative micro-level analyses. While scenes exhibit strong spatial thematic expression, their inherent comprehensiveness can sometimes limit their applicability in analyzing small groups or individual behaviors. In contrast, scenarios, as derivatives within scenes, offer a more nuanced depiction of individual characteristics and play a crucial role in forecasting future trends (Kurniawan and Kundurpi 2019). By distinguishing between collective and individual perspectives, the scenario-based approach provides a more refined framework for exploring the intricate relationships between space and society, addressing the limitations of traditional human geography research at the micro scale. In the AI era, scenarios significantly enhance the diversity and cross-territorial nature of virtual scenes, reinforcing their core analytical value. Individuals from different countries, cultural backgrounds, and identities can now interact within the same virtual scene while engaging in distinct scenarios, thereby positioning scenarios as a critical component in future human geography research that bridges spatial and scene-based analyses.

Overall, Space, Scene, and Scenario each emphasize distinct conceptual domains and analytical dimensions, together constituting a multi-layered pathway from static structures to dynamic processes within human geography. Space refers to the foundational container and field of practice for geographical phenomena, encompassing not only the boundaries of physical space but also the spatial attributes shaped by social, institutional, and cultural forces. Scene represents a thematically defined slice of space—an intermediary analytical unit derived from research-oriented spatial segmentation. Its boundaries are more elastic, characterized by an embedded interplay of function, perception, and sociocultural meaning. In contrast, the scenario dimension focuses on the dynamic processes that unfold within and across scenes, emphasizing their temporal evolution and contextual interrelations. It captures the interactive behaviors and decision-making patterns of specific populations under defined contextual variables, thereby elucidating how spatial practices and societal responses evolve within particular spatiotemporal configurations.

Construction of the 3S research framework in human geography

The 3S research framework in human geography is constructed around three core components: Space, Scene, and Scenario (Fig. 1). It conceptualizes spatiality as a dynamic construct shaped through AI-mediated interactions between real-life and virtual scenes, enabling the generation of knowledge and value across disciplines. Scenarios integrate these layers over time, reflecting evolving, interdisciplinary phenomena grounded in humanities-informed spatial thinking. The three components exist in a mutually dependent and coexistent relationship, without any hierarchical distinction. The construction of the 3S research framework is primarily aimed at addressing human geographical issues in the artificial intelligence era. With the advent of increasingly diverse quantitative data technologies, research in geography has progressively shifted toward element-based data analytics, wherein spatial elements are combined based on data-driven correlations to identify and investigate problems. However, this mode of problem identification risks detaching human geographical analysis from real-world contexts, as it tends to overlook the critical role of human agency and subjective behavior in shaping spatial dynamics.

Fig. 1
figure 1

Space–Scene–Scenario (3S) research framework.

Full size image

In other academic disciplines, space is often conceptualized as a flat, empty physical substrate—a neutral backdrop awaiting occupation or utilization. However, such a notion fails to account for the profound influence space exerts on social life. From the perspective of human geography, space is not merely a selected location for investigation; it constitutes a foundational condition that shapes social structures, human behaviors, administrative arrangements, and trajectories of future development. Therefore, research in human geography must be grounded in the spatiality generated through the attributes and empowerment of space. Spatiality, in this context, does not refer simply to surface characteristics, but rather to the layered meanings inscribed upon space through social discourse, cultural symbolism, historical narratives, and institutional codification. Spatiality is not intrinsic to space itself; it is constructed through sustained practices of classification, naming, boundary-making, and symbolic representation. The study of spatiality thus moves human geographical analysis beyond the visible features of sites or regions, toward an understanding of how space is continually reshaped and re-signified within complex matrices of power relations, identity formation, social inequality, and knowledge production—ultimately revealing the deeper structures and dynamic interrelations between humans and their spatial environments.

Once spatiality is acknowledged, researchers can correspondingly delineate space based on the thematic focus of either real-world or virtual scene. With the advent of the AI era, real-world and virtual scene have entered a state of mutual interactivity (Balsa-Barreiro et al. 2024). Researchers are now able not only to “project” feedback from virtual scenarios onto real-world spaces but also to “simulate” social processes of real-world scene within virtual environments, thereby engendering a bidirectional, coupled mechanism of “hybrid spatial delineation.” This mechanism facilitates a more comprehensive analytical framework for research themes. For instance, in studying urban poverty, impoverished neighborhoods in the physical city can be classified according to socio-economic indicators into low-income or high-risk regions, while concurrently, virtual platforms enable their categorization into perceived poverty, collective anxiety, or platform-based isolation through dimensions such as public sentiment intensity, emotional expression, and social network density.

The final component of the 3S research framework is the concept of “scenario,” which serves as a concrete manifestation of the interplay between human behavior and thematic elements within a given spatial context, thereby illustrating the inherent diversity of scenarios. Within the 3S framework, humans remain the central and unifying element throughout. Different individuals, situated in varying contexts, evolve distinct scenarios shaped by multiple factors such as income, environment, emotional states, and occupational conditions. This centrality of the human subject enriches the 3S framework, underscoring the unique and enduring value of human geography, which resists complete substitution by artificial intelligence.

Within a given thematic scene, some situations align predictably with the scenario’s core theme, while others, despite appearing incongruent, retain an intrinsic logical coherence (Splett 2010). For example, in a scene themed around dining in North America, it is often observed that the primary clientele of Chinese fast-food chains are not necessarily Chinese or Chinese-descendant individuals. This seemingly paradoxical scenario arises because these establishments have localized their menu offerings and business models to better cater to the tastes and lifestyles of the local population. As a result, newly arrived Chinese immigrants or individuals with a strong attachment to traditional Chinese culinary culture may find Westernized Chinese fast food unfamiliar or even unappealing. While this scenario appears to contradict the thematic identity of “Chinese restaurants,” its internal logic remains self-consistent. The dynamic evolution of such scenarios not only influences key aspects of the scene, such as its operational model and level of comfort, but also feeds back into the scene itself, further shaping its trajectory of transformation. Another illustrative case is McDonald’s, a globally recognized fast-food brand that employs localization strategies in different regions. For instance, in India, where religious beliefs largely prohibit beef consumption, McDonald’s offers a menu primarily based on chicken and vegetarian options. In Japan, the company adapts to local tastes by introducing items such as teriyaki burgers and matcha-flavored desserts. Although these adaptations remain within the overarching “McDonald’s” fast-food scene, the specific scenarios exhibit significant variation due to cultural and regional differences.

Scenarios do not exist in isolation but constitute an interactive form of social construction. With the advent of the artificial intelligence era, scenarios across disparate geographical regions can be interconnected and nested through emerging technologies such as augmented reality (AR) and virtual reality. These technologies enable the fusion and reconfiguration of previously discrete scenarios within virtual spaces. For instance, interactive gaming platforms like PlayStation and Nintendo Switch now allow users from diverse geographic locations to simultaneously inhabit a shared virtual environment, wherein individuals generate unique scenario experiences through autonomous configuration and embodied participation. These differentiated scenario trajectories continually intertwine through interaction, culminating in a composite phenomenon that transcends singular geographic locales and linear narrative structures. Over extended temporal spans, this integrative phenomenon synthesizes historical, cultural, and economic elements, which in turn confer attributes and agency to space, thereby enabling the 3S research framework to cohere as a holistic analytical model.

Analytical approaches in the 3S research framework

The 3S framework establishes a robust theoretical foundation for scholars to rethink the evolution of human geography in the AI era. However, its practical implementation requires the integration of theory and methodology through three distinct analytical approaches: the Scenario-Based Approach, the Scene-Based Approach, and the Space-Oriented Approach. By leveraging these perspectives, the framework can effectively address the multifaceted challenges of the future world.

Scenario-based approach

The scenario-based approach represents a relatively traditional method of human geography research (Fig. 2). It typically involves observing specific, smaller groups or individuals. This approach can be categorized into two forms: participatory and non-participatory observation. The subjectivity of the researcher can significantly influence the outcomes of both forms, shaping the interpretation and analysis of the findings.

Fig. 2
figure 2

Scenario-based approach research framework.

Full size image

Participatory observation requires relatively high expertise in human geography from the observer. Observers must make objective judgments based on subjective behavioral observations (Savage 2000). Common methods of participatory observation include field studies, tourism, surveys, and city walks. Long-term local observers (residents), due to their deep familiarity with and affection for local culture, customs, topography, and commercial activities, may inadvertently overlook scenarios of research significance occurring within the scene. Conversely, non-local observers, characterized by heightened sensitivity and curiosity toward unfamiliar environments, often maintain a relatively objective perspective during field visits; however, their observations are susceptible to a range of contingent factors such as the duration of their stay, prevailing weather and climate conditions, and local holidays. When observing and studying scenarios, it is imperative to situate them within a contextually defined scene based on thematic relevance for comprehensive analysis. By comparing the spatial relationship between human behaviors and the thematic elements of the scene, one can construct a dynamic and static synthesis of spatial phenomena. Observers then assess the degree of coordination between people and their surrounding scene and investigate the underlying causes. For phenomena consistent with the thematic scene and exhibiting no discordance, research can focus on elucidating the factors contributing to their effective functioning. In contrast, for discordant phenomena, qualitative analysis should be undertaken by examining points of conflict, integrating theoretical insights from sociology, economics, and psychology, complemented by small-scale data collection to facilitate sample-based interpretation and explanation.

Non-participatory observation (Moug 2007) refers to instances where researchers are unable to conduct direct field visits to target sites and instead rely on secondary sources such as news reports, social media promotions, online video dissemination, third-party narratives, and on-site photographs to observe scenarios. Unlike fieldwork, mediated observation is inherently contingent and opportunistic: observers analyze unfamiliar or unknown scenarios prompted by incidental exposure, critically assessing the credibility and plausibility of the geographic information presented (Yu et al. 2024). To verify the authenticity of such information, some scholars undertake targeted field investigations, thereby transitioning from non-participatory to purposive participatory observation. However, for researchers unable to access the sites physically, it is imperative to rigorously verify information sources or corroborate data through third-party channels. This necessity becomes even more pronounced with the maturation of AI technologies, which demand multi-source validation to ensure the reliability of the research foundation. For example, social media platforms such as Instagram increasingly employ AI-driven image editing or fabricate images tailored to users’ preferences through filters and synthetic technologies. Many composite images alter key elements—such as background colors, local climate representations, and crowd sizes—thereby distorting the scenes and undermining their value for scholarly analysis.

Upon ensuring the accuracy of information obtained through media sources, it becomes essential to identify the geographical area where the scenario occurs and delineate the corresponding research scene in accordance with the specific thematic focus. Given that researchers observe scenarios indirectly via media, lacking sustained, in situ observation of behaviors within the research scene, it is necessary to amass extensive spatial data and select appropriate temporal intervals to compensate for the absence of direct experiential insights into the interplay between themes and behaviors. In non-participatory observation, discrepancies between the scene and human behavior are often discernible to researchers at the point of media information acquisition. Furthermore, non-participatory observation is subject to disciplinary biases; for example, media information, images, and short videos originating from rural contexts frequently receive insufficient attention or interest from urban geographers. In both participatory and non-participatory settings, the 3S research framework serves as a valuable analytical lens that enables researchers to effectively bridge embodied experience and mediated understanding, while preserving methodological consistency and thematic depth across diverse observational conditions.

Scene-based approach

As human society evolves, contemporary research areas increasingly encompass real-life and virtual scenes that are defined by the convergence of multiple and often overlapping thematic trajectories. The selection of these themes shapes the research focus and content directly. A well-chosen theme not only helps concentrate on the research objectives but also guides scholars in identifying and analyzing the functional aspects of associated scenarios. Once a theme is determined, it is essential to integrate the functional characteristics exhibited by related scenarios to then construct a comprehensive phenomenon echoes the theme of the scene (Fig. 3). Specifically, the functionality of these scenarios is reflected in their behaviors and roles under various conditions. For example, while universities are thematically characterized by teaching and research, not all scenarios occurring within university settings are directly related to these core functions. Summer visitors touring the campus, cultural and recreational activities, and student dining gatherings coexist within the same spatial scene as teaching and research activities. Nonetheless, behaviors emerging from these thematically peripheral scenarios also form integral components of the broader spatial phenomena. Through systematic analysis of these behaviors, researchers are able to discern the critical phenomena embedded within a scene that constitute priorities for further empirical and theoretical inquiry. These phenomena, serving as the core of the research, help define the research objectives and guide the selection of relevant factors for deeper analysis. To ensure a holistic and accurate understanding, it is also necessary to consider the conditions of surrounding or similar areas, thereby developing a broader perspective that includes physical, social, and cultural spaces. By synthesizing multiple factors, this approach enables a spatial analysis of scenes, ultimately leading to a comprehensive understanding of spatial scenes and their broader implications.

Fig. 3
figure 3

Scene-based approach research framework.

Full size image

It is important to emphasize that with the continued advancement of AI technologies, tools such as virtual reality and digital twins have begun to offer researchers immersive, realistic, and multi-sensory virtual environments. Following the COVID-19 pandemic, public acceptance of and demand for virtual reality have surged, with devices like Apple’s Apple Vision increasingly entering everyday life. Although current virtual scenarios often lack diverse situational richness, virtual environments constructed from accumulated big data not only provide researchers with accurate models of real-world scenes but also facilitate seamless transitions between similar or adjacent scenarios, thereby reducing the costs associated with physical mobility. Looking ahead, real and virtual environments will coexist in the AI era; their integration into a unified framework will be essential for advancing the study of diverse thematic issues within human geography and spatial research.

Specifically, scene-based research typically entails several critical steps. The first step involves clearly defining the research theme and based on this, identifying appropriate real or virtual scenarios. This process may either start from established spatial contexts to select a research theme or, conversely, proceed from a specific academic question to identify the most representative or explanatory scenarios. The second step involves screening variables of different scenarios according to the types of scenarios encompassed within the thematic focus, thereby enabling a systematic analysis of the dynamic contexts and behavioral logics inherent in scenario research. The third step synthesizes the overlapping mechanisms of various scenario variables within a given context to distill an integrated spatial phenomenon or behavioral pattern, which subsequently serves as the foundation for constructing the research framework. Building upon this, researchers must select relevant social, economic, demographic, and other factors pertinent to the specific research question, employing empirical methods such as statistical modeling, quantitative indicators, or in-depth interviews to conduct analysis. Moreover, in defining spatial scales, a multi-level comparative approach may be incorporated, encompassing the target region, its adjacent neighboring areas, and other comparative regions with similar functions or characteristics, thereby enhancing the explanatory breadth and spatial coherence of the study.

Space-oriented approach

In human geography, spatial research often emphasizes the relationship between space and society. While studies conducted on the macro scale help provide a clear understanding of the geographic spatiality of phenomena, they may still overlook the humanistic aspects. Research in human geography that adopts a space-oriented approach first focuses on the influence of human populations on space. This requires prolonged observation of the study area to reveal the operational mechanisms of the space. Subsequently, phenomena that align with these mechanisms can be identified and analyzed on the macro scale (Fig. 4).

Fig. 4
figure 4

Space-oriented approach research framework.

Full size image

Scholars analyze phenomena such as industries, transportation, and urban clusters within a region to identify the mechanisms driving spatial operations. Phenomena that demonstrate positive impacts within the study area can serve as the foundation for developing a spatial operational model tailored to that region. However, if certain phenomena fail to exhibit functional benefits, human geography must explore related contextual factors to identify the dysfunctional scenes within the spatial model. By analyzing these issues, appropriate optimizations can be proposed to ensure that the spatial operational model aligns with the developmental needs of the region. The inefficiencies in spatial operational models often have multiple causes, but human geography prioritizes identifying the human-related factors. Although studies may overlook individuals or specific groups within the spatial framework, critical mechanisms inevitably leave clues within particular scenarios. These clues, when examined in the context of thematic scenes, can reveal the root cause of the problem. Ultimately, such an approach enables the identification of the core issues and provides insights into how spatial operations can be refined or restructured effectively.

Interdisciplinary applications of the Space–Scene–Scenario (3S) framework in human geography

Spatial disparities and spatial inequalities

The establishment of a research framework aims to identify problems, determine methodologies, explore the value of academic studies, and ultimately lay a foundation for advancing disciplinary theory. In the 20th century, as the study of human–environment relationships in human geography became increasingly mature, spatial disparities emerged as another significant research focus. This concept shifted the emphasis away from the deterministic role of natural environments and toward the diversity and differentiation of human activities, social phenomena, and cultural characteristics within various geographic spaces or regions. The study of spatial disparities offers critical insights into the complex relationships between geographic phenomena and social processes, providing robust support for theoretical development in geography and its practical application in policymaking. In human geography, the term disparity does not have an inherent negative connotation; rather, it is used to describe differences between regions, highlighting variations in natural environments and cultural diversity within spaces. However, in research paradigms increasingly dominated by quantitative methods, disparities have been gradually reframed primarily as value judgments, where differences are assessed as either positive or negative, leading to an oversimplified equation of disparity with inequality (Müller and Ramm, 2024). If human geography focuses on reducing disparities as a means of addressing perceived injustices or inequities within spatial contexts, it may inadvertently undermine spatial diversity and even exacerbate inequalities within these spaces. Therefore, careful consideration must be given to preserving the richness of spatial diversity while addressing spatial inequalities, ensuring that interventions are both equitable and sustainable.

As humanity relies increasingly on technological advancements to overcome the limitations imposed by natural environments, instances of inequality have not diminished but rather, have intensified. Gaps in income levels, access to green living spaces, and available transportation between cities or communities are widening. The apparent cause of this growing disparity lies in national efforts to expand macroeconomic output rapidly by concentrating resources through the construction of urban clusters, leading to the phenomena of urban polarization and collapse. However, the deeper issue stems from the lack of a cohesive theoretical framework within human geography to address and theorize spatial inequality systematically. To address this challenge, the focus of human geography research has shifted from examining spatial disparities rooted in human–environment interactions to exploring ways to resolve spatial inequalities. Studies on spatial inequality over the past two decades can be categorized into three primary approaches: (1) identifying the degree and configuration of spatial inequities through various factors or indicators; (2) reallocating facilities and resources within unequal regions based on elements such as population, income, and transportation; and (3) employing new concepts to redesign urban spaces.

In addressing inequalities among individuals or groups, a common issue—aside from certain studies being disconnected from practical realities—is the high mobility of populations, which often renders research findings obsolete by the time they are applied. Additionally, if resource allocation is based solely on standards derived from conditions at the time of study, this approach risks cementing groups into their existing geographic areas. For instance, using the 15-, 25-, 30-, and X-minute living circle studies as examples (Papadopoulos et al. 2023), designing 15-minute living circles based on the current socioeconomic conditions of residents within specific communities could perpetuate the status quo. Low-income neighborhoods would then remain occupied by low-income groups, ethnically homogeneous communities would stay fixed, and cities would devolve into isolated, enclosed zones. With the widespread adoption of digital technologies and the onset of the AI era, where individuals can increasingly transcend spatial constraints through devices such as smartphones, human geography must leverage the 3S framework to study and address spatial inequality dynamically and adaptably.

Research on spatial inequality within the 3S framework

When spatial inequality is not entirely determined by geographic conditions, exploring and addressing these inequalities through different models becomes a valuable area of study. Achieving absolute equality in human society is inherently challenging, as it involves a range of contested standards including gender, nationality, religion, and income. Human geography offers two fundamental approaches to addressing spatial inequality: spatial resource reallocation (Jiang and Chen 2021) and spatial mechanism reshaping (Magalhães and Carmona 2006) (Table 1).

Table 1 Methods of addressing spatial inequalities.
Full size table

The perspective of spatial resource reallocation posits that spatial inequality often manifests as an uneven distribution of resources, services, and infrastructure (Newman et al. 2021). Under the premise that the existing spatial operation mechanisms are retained or that no better alternatives exist, compensatory measures are implemented to provide residents in different regions with access to comparable resources. This ensures relatively equal opportunities for education, employment, and development across spatial contexts. The academic community has introduced concepts such as Spatial Equilibrium Theory, Spatial-Technology Integration Theory, and Spatial Interaction Models to analyze spatial resource distribution. Additionally, technologies like Geographic Information Systems (GIS) have been employed to systematically synthesize and summarize the allocation mechanisms, patterns, and governing principles of spatial resources. For instance, while developed regions and urban centers may have better schools, healthcare facilities, employment opportunities, and public transportation, underdeveloped areas and rural peripheries often lack these resources. To address spatial inequality, spatial resource reallocation relies on policy interventions minimizing the disparities in opportunities caused by geographical location. The perspective of reshaping spatial mechanisms goes beyond merely addressing the current state of resource distribution; it also emphasizes the pre-existing inequalities in resources and opportunities that arise from spatial differences at the outset. This viewpoint posits that the root cause of spatial inequality lies in the structural constraints imposed by existing spatial mechanisms, and that reshaping these mechanisms is essential for achieving greater equity (Yang and An 2020). In support of this perspective, the academic community has introduced theoretical frameworks such as spatial restructuring, spatial justice, and decentralization to explore strategies for mitigating spatial disparities. In the process of reshaping, it is essential to repair existing spatial operation mechanisms to address the fundamental causes of spatial inequality, thereby enabling people in all regions to start on an equal footing. However, before the advent of AI technologies, reshaping spatial mechanisms was exceedingly challenging. This difficulty arose because information technology still guided population mobility based on spatial agglomeration models, limiting individuals’ and spaces’ options to a significant extent.

The 3S research framework provides a unique perspective and solution to address inequality in human geography at the theoretical level. In the context of spatial resource allocation, the 3S framework identifies significant spatial inequalities in resource distribution effectively across regions and helps define the characteristics of resource-deprived groups. First, the study of spatial inequality must be deeply rooted in the local cultural, institutional, and national contexts to ensure relevance and accuracy. Second, the scenario-based approach ensures that observed spatial inequalities are grounded in the actual circumstances of residents, avoiding resource allocations that do not meet genuine needs. Third, within the integrated attributes of scenes, resource allocation should be based on the overall performance of residents within a given scene, rather than relying on a single indicator to determine distribution standards. The 3S framework emphasizes a dynamic understanding of communities, reflecting the diversity and complexity of residents’ needs across different scenarios. Fourth, in the process of resource allocation at the spatial level, simulations of various scenarios under the 3S framework move beyond simple proportional relationships between factors. Instead, they require the comprehensive consideration of multiple variables and their interactions to develop more precise resource allocation models. The analytical tools provided by the 3S framework not only highlight the outcomes of resource allocation but also offer a robust foundation for formulating effective solutions. Against this backdrop, the concept of smart cities emerges as an ideal model for spatial resource allocation. Smart cities, empowered by AI-driven systems, first integrate and interpret the political, cultural, and historical factors unique to their regions (Zhen and Kong 2021). By “understanding their own unique DNA,” cities can undertake resource allocation tailored to their distinct characteristics. Second, AI systems can simulate personalized scenarios for residents within different urban contexts, enabling more efficient, and equitable resource distribution based on these customized needs. By integrating the 3S research framework with the practice of smart cities, human geography can not only gain a more thorough understanding of the complexities of inequality in the AI era but also leverage AI technologies to propose innovative approaches for future resource allocation policies. This synergy promises to address spatial inequality with unprecedented precision and efficacy, providing actionable insights for equitable urban development.

The 3S research framework plays a pivotal role in reshaping spatial mechanisms by focusing on spatial dimensions as the primary object of observation. It aims to examine existing spatial operational patterns systematically and assess the equity of access for residents in different regions to resources and opportunities through the dynamic performance of specific scenarios. The adaptability of spatial operational mechanisms is often reflected in the economic performance of different regions under similar circumstances. If significant disparities in performance emerge within a region under identical scenarios, leading to systemic spatial inequality between related thematic contexts, the 3S framework traces the formation and roots of these spatial mechanisms to evaluate the potential for optimizing the existing structures. When optimization proves unfeasible, the framework advocates a systemic reshaping of spatial mechanisms aligned with contemporary demands, designing and implementing new mechanisms to address institutionalized inequalities and foster regional equity and coordinated development. This research approach not only prioritizes uncovering the intrinsic logic and evolutionary patterns of spatial mechanisms but also strives to provide theoretical support and practical guidance for modern spatial governance and the realization of social equity. Against this backdrop, urban agglomerations have become key case studies, as they serve as human–geographic units that vividly illustrate the characteristics of spatial operational mechanisms within a region. Historically, urban agglomerations have often adopted single-center or multicenter resource-concentration models to achieve rapid macroeconomic growth. However, these aggregation models have contributed directly to disparities in starting points for residents across different areas within a region, fostering inherent inequalities. The quality of life, access to information, and educational opportunities available to individuals born in central cities have increasingly diverged from those available to residents in peripheral or noncentral areas. In response to the inherent inequalities caused by the spatial mechanisms of urban agglomerations, various nations with such formations have implemented diverse resource allocation strategies. However, these efforts have struggled to mitigate spatial inequalities significantly. Recently, the rapid development of AI technologies has prompted scholars to recognize the potential for directly addressing inequality at the level of spatial mechanisms, bypassing the need for complex and often controversial resource allocation processes. As the spatial mechanisms within urban agglomerations undergo reshaping towards decentralization, hierarchical divisions among cities are gradually dissolved, and metropolitan areas break traditional boundaries to achieve greater spatial integration. This transformation enables individuals within each scene of an urban agglomeration to function independently (Nilforoshan et al., 2023), centering their operations around their own agency, thereby shaping their own scenarios and ultimately achieving equitable starting points. AI technology not only introduces new possibilities for resource allocation within urban agglomerations but also enhances policymakers’ ability to identify and address spatial inequalities more effectively. By employing the 3S research framework, scholars can gain a more comprehensive understanding of the complexities of inequality, leverage the framework to develop appropriate solutions, and propose targeted policy recommendations that foster regional equity and resource sharing across different spatial contexts.

Conclusion and discussion

Discussion

When analyzing complex social phenomena and geographical issues, the Space-Scene-Scenario (3S) research framework demonstrates its unique advantages in adapting to technological advancements and evolving societal contexts. However, as society continues to develop, the 3S framework also encounters a range of challenges and opportunities.

Firstly, this study focuses on the introducing and constructing of the Space-Scene-Scenario (3S) research framework, its analytical approaches, and the relevant theoretical underpinnings, rather than on extensive empirical case studies. In the future, the application and refinement of the 3S framework will rely on researchers integrating it into practical studies. The effective application of the 3S framework requires scholars not only to be well-versed in traditional geographical analysis methods but also to incorporate AI technologies. In particular, the impact of scenario-based interactions derived from virtual scenes on spatial mechanisms will be a crucial direction for future research. Second, the current 3S research framework has been distilled from prior research experience and contemporary societal demands. However, as AI technologies continue to evolve in unprecedented and transformative ways, future applications of the 3S methodology must remain responsive to these advancements. It is essential to ensure that the framework is continually updated and capable of engaging in a constructive and adaptive dialogue with the ongoing development of AI. Thirdly, while the 3S framework offers a multi-dimensional and multi-scalar perspective, its integration with other disciplines—such as sociology, economics, and environmental sciences—remains an unresolved challenge. Future research must explore strategies for fostering deeper interdisciplinary collaborations to expand the scope of the 3S framework and enhance its applicability across diverse research domains. Fourthly, as exemplified by significant policy shifts led by the U.S. government, spatial inequalities across different scenario themes are expected to become more pronounced in the future. The 3S framework must further deepen its analytical capacity to address spatial inequality. Future studies should focus on the interplay between spatiality, scene-based themes, and evolving scenarios, providing a robust theoretical foundation and actionable policy insights for addressing regional disparities in resource allocation and equal opportunity. Finally, the 3S framework offers a novel interdisciplinary approach to human geography. However, as the influence of the social sciences within academia declines and human geography remains fragmented across various theoretical traditions, an essential challenge lies in leveraging the 3S framework to reassert the value of the discipline and attract future scholars. The continued development of the 3S framework hinges on its active application, refinement, and dissemination by scholars across different domains of human geography, ensuring its lasting relevance and significance in the discipline’s evolution.

While the 3S research framework offers valuable conceptual insights, it remains in the early stages of theoretical development and presents several areas for further refinement. First, the framework does not yet provide a systematic approach for integrating interdisciplinary knowledge across fields such as the social sciences, engineering, and the humanities. Its capacity to serve as a bridging methodology across these domains is promising but not yet fully realized. Future work should therefore focus on refining the model’s methodological scaffolding to enhance its relevance and applicability in cross-disciplinary research contexts. Second, the framework has not fully addressed the emerging spatial and perceptual paradigms influenced by artificial intelligence—particularly as experienced by younger generations, such as Generation Z. As AI technologies are reshaping everyday spatial experiences and increasingly mediating new forms of interaction between real and virtual environments, there is a growing need for theoretical models to reflect these shifts. To maintain its theoretical relevance in an era of accelerated digital transformation, the 3S model should further account for these evolving dynamics and incorporate mechanisms to analyze the shifting logic of scene-based interaction shaped by AI technologies.

Conclusion

The core of human geography lies in the value of scenarios created by human activities within spatial and regional divisions defined by specific themes. As AI-driven societal transformations accelerate, disciplines within the humanities and social sciences must anchor their analyses of real-world problems in their fundamental values while actively fostering interdisciplinary integration and collaboration. This principle is particularly crucial for the future development of human geography.

First, the 3S research framework—comprising Space, Scene, and Scenario—explicitly defines the disciplinary value of human geography as residing in the relationship between humans and space. The core objective of human geography extends beyond merely identifying and analyzing the processes through which spatial value is generated; it also seeks to enable individuals to fully realize their humanistic potential within specific spatial contexts. In the AI era, human geography not only distinguishes itself from other social sciences but also integrates insights from fields such as information science, mathematics, and artificial intelligence, transforming these external technologies and methodologies into an organic part of the discipline’s development. This integration of internal disciplinary knowledge with external resources not only enhances human geography’s theoretical explanatory power and practical applicability but also establishes its foundational role in interdisciplinary research. Second, the 3S research framework provides human geography with a novel theoretical tool that aligns with the study of human society in the AI era. Within this framework, Space embodies spatiality, Scene possesses thematic specificity, and Scenario encapsulates creativity. By centering on human activities, the framework organically links these three dimensions, offering a comprehensive analytical structure. Moreover, AI-driven virtual scenarios introduce increasingly complex and diverse influences on spatial and scenario-based studies, further augmenting the explanatory power and practical significance of human geography in the AI era. This framework grants human geographical research a more open perspective and robust methodological support. Third, by constructing a multi-dimensional and multi-scalar analytical structure, the 3S framework effectively circumvents the limitations and rigidity of traditional research approaches. It demonstrates exceptional adaptability and integration across the epistemological systems of human geography and related disciplines, reinforcing the discipline’s ability to remain at the forefront of both theoretical innovation and practical adaptation in a rapidly evolving socio-technical landscape. Finally, spatial inequality will be a pivotal issue in the future development of human geography, with research findings playing a critical role in shaping global development agendas. The 3S framework explores two primary pathways for addressing spatial inequality: (1) leveraging AI-driven algorithms to optimize spatial resource allocation and regulation with greater precision and efficiency, and (2) fundamentally reshaping spatial mechanisms through AI-assisted human intervention. These approaches can be applied across different scales of human geographical inquiry, including urban agglomerations, smart cities, and local living spaces. By incorporating distinct analytical perspectives through Space, Scene, and Scenario, the framework not only provides a novel lens for examining complex socio-spatial inequalities but also establishes a robust theoretical and technical foundation for constructing more equitable and sustainable spatial governance mechanisms.