The wide variety of VGI available today naturally leads to myriad potential GEOINT uses
What is volunteered geographic information (VGI)? Renowned geographer Michael Goodchild coined the term in 2007 in response to the growing phenomenon of laypeople creating and sharing mappable data via the internet. Since that time, geographers have debated the scope of the term and how it differs from related concepts, such as public participation GIS, which existed prior to Web 2.0 and its concomitant collection of user-generated content. VGI is a broad term, and one way to categorize types of VGI is to consider the intentionality of the volunteer. From most to least intentionality, we might identify five categories/levels of VGI: 1) a volunteer who is actively involved in framing the project; 2) a volunteer who helps devise the goals and collect data, but does not analyze data or make scientific findings; 3) a volunteer who knowingly contributes VGI for a particular purpose; 4) a person who contributes VGI for a different purpose; and 5) a person who unwittingly creates VGI.
Examples include persons involved in citizen science projects (Levels 1 and 2), OpenStreetMap users (Level 3), a Facebook user who knowingly shares a geotagged photo which is then downloaded and shared widely (Level 4), and a person who has unwittingly enabled cell-phone tracking on various smartphone apps (Level 5). Levels 4 and 5 might be considered contributed geographic information (CGI) or ambient geographic information (AGI), rather than true VGI. In such cases, while the contribution may technically be voluntary, the information is volunteered for a different purpose than it might ultimately be used or even unknowingly.
Broadening the discussion through further demonstrations, an example of VGI at the most collaborative levels is suggested by Dr. Mordechai Haklay, a leading researcher of GIS and citizen science: Professional and non-professional astronomers often work together on projects that are integrated at all levels, including the definition of the scientific question, field collection methods, and analysis. Haklay suggests that environmental justice projects often involve VGI at slightly less collaboration, wherein community residents work with environmental scientists to define the project goal and to collect data, but do not perform scientific analysis.
Another step down is seen in Dr. Elisabeth Sedano’s study of outdoor advertising in Los Angeles, wherein residents helped to map billboards using a web map for a project whose goals and methods were shared with the volunteers but were defined by the researcher. The lowest level of collaboration in VGI is seen in projects based on geotagged social media posts, such as tweets, as users have voluntarily shared their location but without the specific intent to add to any project beyond their chosen social media platform.
Because there is no intentional collaboration, Dr. Anthony Stefanidis of George Mason University suggests such “volunteered” data should be characterized as ambient geographic information (AGI) rather than true VGI. In other cases, users may not realize an app is set to record their location, or their location contribution may technically be voluntary, but was volunteered only as a means to be able to participate in a particular web-based activity, where consenting to provide user location, for example, is a requirement to proceed. These categorizations are important to understand; ethically, they help define the intended use of the information as volunteered by the provider.
Current Thinking on the Ethical Use of VGI
With respect to VGI contributors, the discussion of ethics generally focuses on privacy. Regarding the aggregator and distributor of VGI, ethics might also involve legal liability, responsible use, and data quality considerations. Studies of VGI often focus on social media posts during disaster and crisis events, in which data accuracy can have vital consequences, whereas private sector data quality is less an issue than invasion of privacy. In the commercial sector, outdoor advertisers are combining location-based services, big data on consumer spending patterns, and digital sign displays to make real-time changes in advertisements based on nearby individuals. One of the experts in the field of VGI is Dr. Peter Mooney of Maynooth University, Ireland. He states that private data in the VGI context are often any geographic data or information that can be linked to an individual contributor who created, collected, or edited it. A key aspect of project planning therefore is to consider whether data to be collected needs to be linked to its creator for the purposes of the project. Linkable data should only be collected if the project depends upon it and with the consent of the creator. When linkable private data are collected, protections should be established to ensure the data are only used according to the purpose for which the creator consented or to take active steps to de-link data if linking is no longer required. If linkable private data are collected, it then becomes necessary to establish some protection that ensures the data are only used according to the original purpose defined before the collection of the VGI started. When VGI data collections are considered a resource for new and maybe unforeseen beneficial uses and research, it is arguably more important that these data do not provide linkable private data about the individuals who contributed it.
In a geospatial intelligence (GEOINT) context, the question that should be asked is whether location information in itself is private data or can be linked to individuals. The answer often then depends on the location accuracy. Many location data are accurate enough to be connected to one individual or to a small group of individuals (like an office or home). Sometimes this information is even combined with precise time and date. Mooney notes that there is no common solution; rather, the collection of point-based geographic data for a specific purpose may need to have high geographic accuracy which produces a possibility that the geographic features close to the collected points could be used to infer other information. Supporting Mooney’s work, Barbara Poore, a research geographer from the United States Geological Survey, notes that the possibilities surrounding the use of location data associated with an identifiable contributor or embedded GPS information from cameras might entice an unethical use of the VGI. Finally, recognized methods to protect or hide information fidelity of data exist, such as making location information blurry or fuzzy and anonymizing the data by making private information available only per contributor’s preferences.
- This article is part of USGIF’s 2018 State & Future of GEOINT Report. Download the PDF to view the report in its entirety and to read this article with citations.
Possible GEOINT VGI Use
The wide variety of VGI available today naturally leads to a myriad of potential uses for VGI as GEOINT. The use of VGI can be leveraged to significantly enhance conventional intelligence capabilities for Department of Defense (DoD) and homeland security organizations. The upsurge of web-based technologies that allow individuals to voluntarily develop applications and provide information offer numerous opportunities to improve GEOINT collection, management, retrieval, and dissemination. VGI technologies currently being developed and used by the mainstream population can likely be adapted for use with multi-echelon security access. In this, the use of VGI type sites as a structure for collection, management, retrieval, and dissemination of classified geographic information could provide the defense and intelligence communities with quickly evolving and continually improving technologies.
Recent lessons learned from operations in conflict areas such as Afghanistan and Iraq have demonstrated a need for improvements in GEOINT collection methods. VGI is a potential mechanism for increasing the number of sensors populating information databases. In this context, every person on the battlefield (including soldiers) is a sensor; therefore, all those on today’s and tomorrow’s battlefields are possible GEOINT providers. Careful investigation, however, into potential incentives for volunteering information is necessary to fully understand the quality of the data provided. The management of intelligence using newly developed technologies and methodologies is significantly different from conventional schemes. In this, a number of questions arise:
- Should databases be populated without restriction and how is the accuracy of the information verified? Contributors may interpret the world differently. For example, what someone calls a bus terminal may be a bus station, stop, or bench to someone else.
- Is a database manager needed to supervise the data published? Some algorithms for data validation might not be able to reach common sense conclusions that a human monitor would be capable of reaching. New technologies now allow the users to become database managers. Websites such as eBay and Amazon allow users to rate each other. Wikimapia allows users to change previously submitted VGI.
- Is it possible for a VGI site to be self-correcting, self-improving, and self-assessing in order to continually judge the quality of the information? A meticulous study of the supervision methods for the provided information is required. In this, data comprehensiveness for VGI should be considered, as many people are not volunteering their information, which could lead to bias in the data or skewed datasets if some users submit multiple VGI. In some cases, a person’s use of multiple devices could lead to the impression that the VGI is from multiple people.
LTC Ian Irmischer, Ph.D., from the United States Military Academy at West Point, noted that from these questions, it appears that GEOINT searches inherently arise from the possibilities of vast information collection. A powerful geospatial search engine that appropriately prioritizes information is essential for the efficient use of VGI. Recognizing that national security users often need geospatial information for immediate response situations, the search engine should be able to conduct network analysis of requested information and analyze the spatial components of the data. Standardization of automated metadata inclusion is likely required to allow the users to query and access needed intelligence. As a result, the compiled VGI must be able to be disseminated and visualized by the user facilitating interoperability between provided data formats and a common operating platform that can be efficiently interfaced by sensors and operators.
Defense and intelligence users would require limited training if currently existing VGI collection methods were integrated. Sites such as Google Earth and Wikimapia are intuitively designed, have widespread use, and are familiar to citizen sensors, organizational sensors, and operators in need of information. Resultant from all of this, VGI has many compelling uses for operators that require GEOINT for situational awareness. The use of citizens as sensors vastly increases the possible GEOINT collection capabilities by governmental organizations. A detailed examination of how VGI and associated technologies can improve the collection, management, retrieval, and dissemination for these organizations could advance local and national level crisis reaction and security.
The National Geospatial-Intelligence Agency (NGA) has progressed quickly into the VGI space through its creation of the NSG Open Mapping Enclave (NOME). NOME is an online tool kit that allows the National System for Geospatial Intelligence (NSG) to contribute and benefit from the power of open content. As a crowdsourcing method to harness tools to create, assemble, and disseminate geographic data provided voluntarily by individuals, crowdsourcing products such as Wikipedia and OpenStreetMap collect useful geospatial information for NGA analysts to produce GEOINT products. By allowing users to contribute their expert knowledge to maps, VGI opens up geospatial data to communities that in the past would rely on their own limited collection resources and proprietary technologies. NOME challenges conventional geospatial collection and dissemination methods to reduce costs, improve accuracy, and enhance mission planning and execution.
The use of VGI for GEOINT is also relevant to humanitarian assistance and disaster management operations and functions. In his work on crowdsourcing VGI, Flávio Horita concluded that the scientific literature about the use of VGI in disaster management is growing, with a significant increase in the number of publications since 2010, noting that the predominant research area was disaster response. Fewer studies were devoted to mitigation and preparedness, while none dealt with recovery. His explanation to this was that response is the most visible part of disaster management and is also more likely to attract the attention of volunteers. However, the challenge for researchers remains how to advance knowledge about methods that include VGI in mitigation and preparedness activities such as risk analysis and early warning systems, as well as in the recovery phase, by helping communities reorganize their routine and create mechanisms to prevent disasters from happening in the future. Horita continued, “This research also showed that VGI is commonly used to manage floods and fires. The prevailing media for sharing VGI was found to be social media (i.e., Twitter, Facebook, YouTube, etc.) and mobile devices. Interestingly, very few of the reviewed papers address VGI platforms like Ushahidi2, Elva 3, OpenStreetMap, and Wikimapia.”
Ethical Use of VGI for GEOINT: Do the Rules Change?
There are ongoing discussions about the ethical use of VGI for GEOINT and whether the standard ethical considerations for VGI use apply. There is concern for targeted data analysis and use, and there appears to be potential for both positive and negative impact. Four areas of concern are noteworthy:
- Data exploitation methods (including commercial, online, and offline use cases).
- Relevant, geographic scale.
- Bounded areas of data exploitation (geo-fences, group audience demographics, etc.).
- Individual protection considerations.
Stefanidis noted in his GeoJournal article that harvesting VGI and ambient information brings forward novel challenges to the issue of privacy, as analysis can reveal information the contributor did not explicitly communicate. Stefanidis said this is not a new trend. For example, Google uses the information it collects to improve its customer service. Similarly, Twitter makes money by licensing a tweet gateway to search engines, while companies can pay for “promoted tweets.” This trend has already spread to location information. For example, TomTom uses passively sensed data to help law enforcement determine the placements of speed cameras. iPhones store location data of which the user may be unaware. Stefanidis said the public is making progress in highlighting the issue of privacy-relinquishing when sharing location information. Sites and apps such as pleaserobme.com and Creepy (a geolocation aggregator) have demonstrated the potential for aggregating social media to pinpoint user locations. In this, trying to protect people’s identities in times of unrest is also a well-recognized concern. For example, the Standby Task Force suggests ways of limiting exposure and delaying information for the recent unrest in North Africa. Stefanidis further stated, “But the power of harvest AGI stems from gaining a deeper understanding of groups rather than looking at specific individuals. As the popularity of social media is growing exponentially, we are presented with unique opportunities to identify and understand information dissemination mechanisms and patterns of activity in both the geographical and social dimensions, allowing us to optimize responses to specific events, while the identification of hot spot emergence helps us allocate resources to meet forthcoming needs.”
Although ethical concerns exist with the use of VGI for GEOINT, we posit some recommendations for going forward: 1) security and safety of the collective takes precedence; 2) personal information may and should be leveraged to ensure the security and safety of said person; 3) accepted, generalized, targeted information passing (aka advertising) should be used as a starting point when communicating general alert notifications; 4) VGI identification can and should be used for general (non-targeted) reverse information passing; 5) more descriptive VGI identification can and should be used for targeted information passing in response to a disaster/emergency/conflict; and 6) caution should be used when using VGI identification for psychological operations in a mission plan (often connected to some form of a military, operational response).
Inherent in the recommendations is that security is a vulnerability, especially with advances in technical capabilities and increasing sophistication of sponsoring enterprises (criminal, opposing government, or otherwise). Security breaches are becoming more common (or more commonly reported), whether accidental (such as a lost, unencrypted laptop or careless disposal of old equipment) or intentional attack. At some point, security breaches are a possibility. In not knowing how the compromised data will be used, unpredictable consequences can ensue. From this and in concert with the recommendations above, targeted advertising and information passing (from the third recommendation above) represents a base level in which no personal information about an individual is necessary. More descriptive VGI identification (from the fifth recommendation above) is recommended when safety is paramount and response time could be a life or death situation; the more information available to assist in evacuation and rescue efforts, the better. In this, one must keep in mind that first responders (in the field) may not have time to sift through data. Someone offsite (or resident within an app) would likely need to filter the information down to the necessary components for the responder to act quickly. More information can sometimes be too much information, depending on the context. Finally (from the sixth recommendation above), in conflict or psychological operations, deadly consequences could result if the opposing force acquires the information, whether or not the VGI aggregator/distributor is aware of the obtainment.