# Module 8. GIScience Ethics ### Learning Objectives * Describe how location accuracy can be modified to increase privacy. * Consider alternative perspectives on ethical considerations about data, privacy, and impact. ### Assignments * Quiz 8 * Lab 8 (Discussion) * Complete your final projects by Friday ### Overview ## Ethics in the Era of Big Data Geospatial data analytics can provide valuable insights and inform decision-making in various domains. However, the use of these methods also poses potential risks and challenges, especially when it comes to privacy, security, and equity. Ethical behavior is considered behavior desired by society beyond the minimum behavior established by law \[1]. It is important to approach the implementation of geospatial data analytics with caution. This includes considering the potential impacts of data collection and analysis on marginalized communities and ensuring transparency and accountability in all geospatial work. This week we will look at several examples where these concerns take center stage. ### Privacy The term Volunteered Geographic Information (VGI) was coined in 2007 by Michael Goodchild \[2]. The term was used to describe activities surrounding tools like Wikimapia and OpenStreetMap. In this paper, Goodchild also alluded to the use of human sensors for VGI, highlighting citizen science applications as one area where VGI had proven useful. Since 2007, there has been an explosion in the number of ways we use humans as sensors. Social media, such as Twitter and Instagram, are common sources of information on human sentiment about a place (see Camacho, et al. for an example on Twitter. Other apps, like Strava or Uber, accurately track our locations and travel patterns. A 2006 Nature article describes an incident where a mortality map created during disaster response to Hurricane was digitized using GIS, pinpointing and making public locations where people’s bodies had been found \[3]. In another example from 2018, the fitness tracking company Strava released a global heat map of its users' activity data, including locations and routes of popular running and cycling routes. The map unintentionally revealed sensitive military bases and patrol routes in countries such as Syria, Iraq, and Afghanistan, as soldiers and military personnel also used the app to track their workouts. The incident raised concerns about the potential risks of using fitness-tracking apps that collect and share users' location data. It highlighted the need for stricter privacy regulations and better data protection measures for individuals and organizations, particularly in the context of sensitive and confidential information. One way to mitigate geolocation privacy concerns is to aggregate or add spatial randomness to the data. The US Forest Service’s Forest Inventory and Analysis program uses both data aggregation and spatial anonymization to protect the privacy of landowners, as described in their documentation: “Actual plot coordinates cannot be released because of a Privacy provision enacted by Congress in the Food Security Act of 1985. Therefore, this attribute is approximately +/- 1 mile. Most plots are within +/- ½ mile. There is additional uncertainty for private plots caused by swapping plot coordinates for up to 20 percent of the plots.”\[4] ### Algorithms In addition to concerns about location privacy, the development and implementation of machine learning algorithms. You may have heard the term “Garbage In, Garbage Out” before. This term refers to the fact that an algorithm is only as good as its data. Similarly, an algorithm is only as ethical as the data it is trained with. Algorithms can perpetuate or amplify existing biases and inequalities if the data used is biased in some way. A common example of this is incomplete or unbalanced sample data when performing classification tasks. Outside of the geospatial domain, you may have encountered examples like the racial biases that arise in facial recognition algorithms due to a lack of representation of minorities in training data \[5]. Classification, as we learned earlier this term, is a common geospatial task, and the completeness and representativeness of the data we use in classification directly impact the results we achieve. Defining classes for classification is a great example of a process where generalization or exclusion could lead to unintended consequences. Global classification schemas, like those used for worldwide land cover or land use classification, are necessarily generalized due to the scale of analysis. However, such generalizations also neglect to represent unique environments for the sake of generalizability. For example, the (FAO) points out that the UNESCO Vegetation Classification only considers only natural vegetation, while all other vegetated areas, such as cultivated areas and urban vegetated areas, are ignored (). Such generalizations could erase land cover changes indicative of unethical land management practices or inequitable land use. ### Data Ownership & Access By now, you have had some experience obtaining and working with GIS data. Some of that data was provided via government web portals like the Census and Earth Explorer websites. Globally speaking, this open access is more of an exception than a rule. Even in the United States, access to our own geospatial data is held behind a paywall. Commercial operations, like Airbnb, do not find data sharing in their best interest, and instead, analysts have come up with different ways to hack their way to the data they want. Airbnb presents a complex situation. As noted by the website Inside Airbnb (), the company has disrupted housing markets and shelter accessibility around the world with its business model. Understanding the impact of the company on local housing markets requires an understanding of who owns AirBNBs, how much are they charging, and where they are. Unfortunately, that also brings up the question of how much privacy are AirBNB owners entitled to. ### Ethical Standards and Practices Many organizations and funding agencies have established guidelines for scientists in ethical data handling. For example, \[6] outlines considerations to make when performing geospatial analyses and provides scientists with a clear checklist of questions to ask in the design and implementation of geospatial research. However, as \[7] suggests, many guidelines only go so far as suggesting ethical principles, including transparency, justice, non-maleficence, responsibility, accountability, privacy, safety, and trust, while falling short of providing guidance on how to achieve it. \ In this module, we took a cursory look at some of the ethical concerns that arise when working with geospatial data. There are numerous other ethical considerations, and I highly recommend you spend some time looking at the areas of critical GIS and critical cartography to explore both research and activism in this area. ### Citations (You do not need to read these) \[1]Onsrud, Harlan J. "Identifying unethical conduct in the use of GIS." Cartography and Geographic Information Systems 22.1 (1995): 90-97. \[2] Goodchild, M.F. Citizens as sensors: the world of volunteered geography. GeoJournal 69, 211–221 (2007). \[3] Curtis, A., Mills, J. & Leitner, M. Keeping an eye on privacy issues with geospatial data. Nature 441, 150 (2006). \[4] Woudenberg, Sharon W., et al. The Forest Inventory and Analysis Database: Database description and users manual version 4.0 for Phase 2. Fort Collins, Colorado, USA: United States Department of Agriculture, Forest Service, Rocky Mountain Research Station, 2010. \[5] Van Noorden, Richard. "The ethical questions that haunt facial-recognition research." Nature 587.7834 (2020): 354-359. \[6] Berman, Gabrielle, Sara de la Rosa, and Tanya Accone. "Ethical considerations when using geospatial technologies for evidence generation." (2018). \[7] Jobin, Anna, Marcello Ienca, and Effy Vayena. "The global landscape of AI ethics guidelines." Nature Machine Intelligence 1.9 (2019): 389-399. ### Assigned Readings * Flowers, R. *Is the debate over the ethical use of geospatial data dead?* Directions Magazine. June 17, 2021 (Accessed: March 4, 2023)