In-Between the Lines and Pixels: Cartography’s Transition from Tool of the State to Humanitarian Mapping of Deprived Urban Areas

Abstract

Cartography has been, in its pre-modern and modern production of maps, influential in determining how space and territory is experienced and defined. But advancements in telecommunications and geovisualization software, along with geoinformation systems and geoinformation science (GIS), have transformed cartographic practice from a tool of dominantly state apparatus to a scientific, commercial, and humanitarian enterprise. This is exemplified in the use of remote sensing (RS) techniques to acquire, process, and visualize images of the Earth. In the last decade, RS techniques have increasingly incorporated Artificial Intelligence (e.g., Convolutional Neural Networks) to improve the speed and accuracy of feature extraction and classification in remotely sensed images. This paper will investigate the use of CNNs in the classification of deprived urban areas referred to as “slums” and “informal settlements” in the Global South. Using a postphenomenological methodology, this paper shall analyze the role of classification and use of geoinformation in shaping how deprived urban areas are algorithmically classified. This analysis will reveal that besides the technical opportunities and challenges, attention needs to be given to three ethical areas of concern: how deprived area mapping using AI impacts the agency of communities, how there is a potential lack in the democratization of these RS technologies, and how the privacy and data protection of communities being mapped is endangered.

Crossing Boundaries: The Ethics of AI and Geographic Information Technologies

“…The data used to produce maps and statistics on the growth of DUAs are gathered from censuses and national household surveys at the municipal level as well as from non-governmental organisations, such as the WHO and UNICEF Joint Monitoring Programme on Water Supply and Sanitation (JMP) [13]. These mapping initiatives have been supplemented by deploying AI, specifically machine learning (ML) based methods such as random forests, support vector machines and deep learning [14][15][16][17]. These ML methods transform spatial data "into different layers of abstraction" that can become useful for making predictions as well as data representations [18].…”

“…Accountability, fairness, privacy and transparency were mentioned in all 16 guidelines, showing the importance given to these values/principles. Trust (15), rights (15), bias (14), discrimination (15), access (14), risk (14), harm (14), responsibility (14), autonomy (13) and inclusion ( 13) also appeared in many of the guidelines. Beneficence (3) was mentioned the least.…”

Data Is Power: Addressing the Power Imbalance Around Community Data with the Open-Access Data4HumanRights Curriculum

“…However, open data to monitor progress towards this target is often unavailable at the required scale and semantic details to support local action [1]. Data on deprived communities is commonly collected by external actors with limited empowerment of community groups [2,3]. Consequently, collected (extracted) data is not available to communities [4].…”

“…While in many contexts, community groups have existing capacities in data collection [13,14], previous research and initial discussions with community leaders demonstrated a massive capacity gap in effectively utilizing community-collected data -specifically to organise, analyse, visualise and communicate data [15]. In many Low-and Middle-income Countries (LMICs), urban deprived communities are exposed to harsh and routine human rights (Human rights-related issues include, e.g., the Right to life, Right to health, Right to education, Right to adequate housing, Right to dignity, Right to personal liberty, Right to fair hearing, Right to freedom of thought, Right to freedom of expression and the press, Right to freedom from discrimination [3]) violations in the form of forced evictions, lack of access to basic services, mass arrests, and police brutality, to name a few. Residents of deprived communities are often stigmatized on account of their poverty [16].…”