By Peter Muskett
A crucial component for analyses within the field of big data analytics is the inherent connection data has with geography and location. [BA1] A commonly cited adage proposes “that at least 80% of all data are geographic in nature” either through the provision of imagery or the ability for data to be quickly georeferenced. Aggregating and interpreting this geospatial data as a means to “empower understanding, insight, decision-making, and prediction” is a core feature of GIS and for years has served as a driving force behind both governmental and corporate processes of operation. Utilizing GIS to build and implement analyses of Location Intelligence is not a new concept; however, the manner and rate at which these existing practices have evolved and paired with new systems of AI and machine learning to build Geospatial Artificial Intelligence (geoAI) have been particularly notable. The deployment of these technologies has the potential to profoundly impact the human condition, therefore there is a growing concern as to how geospatial data can be responsibly collected while still protecting the privacy and overall well-being of the populations being analyzed.
To best answer this question, it is first necessary to examine the benefits of expanding geoAI capabilities to understand why the field has seen an increase in investment in recent years. Due to its interdisciplinarity, the field of geoAI has seen a flexible and diverse customer base utilize its capabilities for a variety of purposes. In business, many companies view investment in geoAI as a matter of survival in a constantly evolving marketplace. Manager of General Motors Advanced Data Analytics, Bruce Wong, for example, states that a pressing concern in the market is that a considerable portion of consumers today find it both financially and psychologically easier to search for and purchase goods through the internet or local online marketplaces rather than through the more traditional means of going to a physical location such as a dealership. As a result, the explosion of commerce conducted through the internet has played a substantial role in pushing companies such as GM to bolster their geospatial artificial intelligence capabilities as a means to strategically place dealership locations to maximize profit in areas likely to see substantial payoff.
From a standpoint of public safety, law enforcement departments have begun heavily integrating geoAI to improve policing and respond more effectively to crime. Methods to accomplish this task range from creation of comprehensive crime maps to the real time analysis of video camera feeds that can recognize “aggressive behavior” and “biometric features.” As police departments in the United States face increasing societal pressure to alter their operations to better serve their communities it will be worthwhile to track into the near future whether this substantial investment into geoAI capabilities will not only decrease crime rates but also improve cultural perception and legitimacy through better policing. Ultimately, this serves to illustrate the rising prevalence of geoAI as a crucial measure for the private industry to remain competitive and relevant to consumers seeking the most optimal and convenient means to purchase, and for governing bodies through the utilization of new technologies to ensure order and safety.
What makes this pairing between remote imaging, georeferencing, and artificial intelligence as a driving force behind profit and public safety particularly concerning, however, is its potential for deepening surveillance practices and risks of data breaches. In the case of remote viewing through the usage of drones or satellites there is almost no possibility for individuals to provide “informed consent” for the monitoring of their private property, if they are even aware of the surveillance in the first place. For this reason, many question the ethics of analyses deriving from information taken without a citizen’s explicit awareness. This also does not begin to take into consideration the detrimental psychological effects and negative public perception that the utilization of UAVs can have on a population. Studies have demonstrated that those, especially who have experienced armed conflict or life in an authoritarian regime, closely link unmanned aircraft with “evidence of spying” and “collaboration” with oppressive governing bodies. Therefore, the presence of such technology, even if used for simple purposes such as detecting land change, can cause distress and mistrust within a population. Additionally, a growing worry regarding the collection of data of this nature is the increasing ease of de-anonymization of publicly available records by linking “data back to individuals using their geographical location.” This concern is compounded by the fact that “comprehensive privacy, data protection and storage standards may be largely non-existent” in states across the globe, highlighting how in certain regions the technology has far outpaced political efforts to responsibly collect, de-identify, and share personal data of citizens.
What has become increasingly clear is that the technologies utilized in geoAI are evolving at a rapid rate and the concerns regarding privacy and general well-being will only worsen if proper action is not taken in the present. Developing technologies such as hyperspectral imaging, which utilizes hundreds to thousands of wavelengths in an image capturing process at the pixel level in comparison with the three to ten bands employed by traditional multispectral imaging, highlight a future characterized by an enhanced ability to quickly and remotely tell a much deeper story about the landscape, or lifeforms, captured than ever before. Already, studies of hyperspectral imaging in a human context have demonstrated the ability to remotely detect and classify human emotion, such as happiness, based on facial expression and tissue oxygen saturation, and disease diagnosis by remotely differentiating between healthy and malignant tissue in a patient. While developing techniques utilized in geoAI such as these present a new and exciting future that can greatly improve the lives and of many, others maintain a justifiable fear of the possibilities if this data is deployed by actors with malicious intent. Additionally, even if used by actors of public trust there exists the question as to whether it is being used responsibly and properly to ensure that uses of the data are not inaccurate or misleading.
At present, a massive and instantaneous collection and analysis of human sentiment and physical health through remote sensing and AI is still a distant prospect. Presently, there are a multitude of barriers ranging from the real-time analysis of big data, to potential privacy concerns over widespread data collection that might deter implementation. At the very least, however, it serves as a relevant example of how geoAI might contribute both to understanding the human condition and uncovering previously hidden patterns, while also raising legitimate concerns of scope and availability. Of particular concern is the utilization of geoAI by government and corporations absent of sufficient legal or regulatory oversight.