Ethics & Privacy Considerations

The application of machine learning to predict wildfires raises several ethical concerns that must be carefully addressed:

1. Privacy Invasion: The growing accessibility of high-resolution satellite imagery has raised concerns about privacy intrusion. Satellites in orbit can capture incredibly detailed views of homes and properties, often to the extent of distinguishing individual objects, vehicles, or even people. This level of detail, when improperly used, can lead to an inadvertent invasion of personal privacy. Concerns may include the ability to identify individuals or track their activities, and the exposure of sensitive personal information such as property layouts or daily routines. The lack of consent and control over what is captured by these satellites underscores the ethical and legal implications of this technology.
2. Data Security: Satellite data is highly valuable, and its collection, transmission, and storage present a set of security challenges. Mishandling satellite data can have far-reaching consequences. Security risks are numerous and may include data breaches where unauthorized parties gain access to sensitive data, which can compromise privacy, national security, or corporate interests. Data manipulation is another concern, as malicious actors could alter satellite data, leading to misinformation or causing serious accidents, especially in sectors like navigation or aviation. Additionally, cyber threats targeting satellite systems can disrupt critical services, such as weather forecasting, communications, or GPS navigation, with potential global repercussions.
3. Bias in Decision-Making: The use of satellite data in various decision-making processes, such as insurance pricing, urban planning, or policy formulation, can introduce inherent biases. This bias can manifest in several ways, impacting fairness and accuracy. For instance, unequal access to satellite technology can result in disadvantaged groups being excluded from benefits or protections based on this data. Historical biases in data collection may lead to the perpetuation of inequalities, affecting insurance premiums, disaster response, or resource allocation. Moreover, gaps in satellite data coverage can lead to underrepresentation and inadequate decision-making for regions or demographics not adequately included in the data.
4. Algorithmic Bias: The algorithms used to analyze and interpret satellite data are not immune to bias. If these algorithms are biased, they can perpetuate discrimination, environmental injustices, and resource misallocation. Biased algorithms may lead to unfair targeting of specific groups for surveillance or law enforcement, potentially violating civil rights. They may also generate incorrect conclusions about environmental concerns, disproportionately affecting marginalized communities. Additionally, resource allocation, particularly in disaster response, can be unfairly distributed due to algorithmic bias, leaving vulnerable communities at greater risk, which raises ethical and social equity concerns.

We've teamed up with a DATASCI 231 Data Privacy student, Edwin Figueroa, to address the ethical and privacy aspects of our project. We share project details and engage in ongoing communication to provide a comprehensive understanding of our objectives and data privacy challenges. At the end of our collaborative efforts, our partner will offer suggestions for us to review. We can then evaluate and decide how best to incorporate these valuable insights into our project, ensuring it aligns with ethical standards and privacy regulations. This partnership underscores our shared commitment to data privacy and offers a valuable opportunity for mutual learning.