PROJECT BRIEF

BACKGROUND

The city of Tempe, Arizona hosts an open-by-default data classification system, which publishes data collected by the government to a database accessible to the public. This has created strong civic transparency and trust among Tempe residents. However, it carries some limitations due to its lack of differentiation among data types. A more nuanced model is needed to maximize internal data utility while protecting individual privacy.

AIMS

This framework seeks to provide clarity in dealing with the data types that do not fall into static categories and create a more sophisticated method for handling data collected by the government. Current data aggregation practices are also reviewed, as artificial intelligence poses increasing risks by allowing for the combination of different data sets to pinpoint individuals.

METHODOLOGY

Project methodology included conducting a comprehensive review of privacy approaches taken by other cities in situations similar to Tempe. Fellows examined systems used in Washington, D.C. which also operates under an open-by-default principle, Seattle, Boise, Chicago, Austin, and Amsterdam, which has been pioneering algorithm classification. Fellows also analyzed the demographic landscape of Tempe, considered government interests, and interviewed Tempe’s Chief Diversity Officer in order to gain a more complete understanding of the stakeholders involved in government data collection. 

DELIVERABLES

Project research suggests the implementation of a two-part system: a value-point strategy layered with a tier-based classification approach that employs various de-identification techniques. Fellows also identified the categories of AI privacy risk relevant to Tempe’s data holdings, drawing on documented cases, and then outlined potential prevention measures scaled to sensitivity tiers established in the classification framework. Lastly, the team applied this data framework to a specific case study, homelessness data collection.

IMPACT AND FUTURE WORK

This project provides a starting platform to help the city of Tempe apply findings to their determined standards and nuanced user needs. This research is intended to protect Tempe’s vulnerable populations, provide ideas to combat emerging risk, and create more varied privacy categorizations and anonymization methods. Additionally, this framework can inspire future open-by-default municipalities, filling the gap for a nuanced privacy structure.