Contextual Integrity for Computer Systems - January 2023
PI(s), Co-PI(s), Researchers: Michael Tschantz (ICSI), Helen Nissenbaum (Cornell Tech)
HARD PROBLEM(S) ADDRESSED
Scalability and Composability, Policy-Governed Secure Collaboration
PUBLICATIONS
KEY HIGHLIGHTS
We continued to apply CI to systematically analyze NIST's "PRAM" (short for "Privacy Risk Assessment Methodology") and more generally our work on privacy risk assessments. We have continued our work on the study of privacy risk assessments' and their role in systems' design. In our previous analysis we had relied on contextual integrity as a theoretical framework to reveal the presence of critical flaws in these approaches. These findings led us to argue that contextual integrity is useful not only to reveal the shortcomings of privacy risk assessments, but also to practically address these shortcomings. More recently, we have extended our analysis to illustrate how the contextual integrity framework can help scope, guide and inform the assessment of privacy risk, as part of an article we are preparing for submission at the Computers, Privacy and Data Protection Conference (held in Brussels, Belgium, in May 2023).
Secondly, we have continued our research on the relationship between differential privacy and contextual integrity. In spite of its apparent properties and promises, differential privacy remains hard to interpret and accommodate into existing data analysis practices and processes. Previously, we had identified three common misconceptions about differential privacy, and how these misconceptions may contribute to legitimize privacy-invasive systems or applications. We have subjected these misconceptions to a contextual integrity analysis to elucidate and better explain how differential privacy is misunderstood, its properties and guarantees misconstrued. This led us to propose more precise modeling and communication methods to separate the narrow notion of privacy that differential privacy represents from the wider, more capacious notion that contextual integrity encompasses, a task for which contextual integrity itself holds promise. Beyond showing the limitations and misconceptions of differential privacy, contextual integrity can help us reason in situations when the use of differential privacy is appropriate and beneficial. Currently we are working on use cases that showcase this interplay. We also plan submit an extended abstract at the Privacy Law Scholars Conference, to be held on June 1-2 in Boulder, CO.
We also continued working on how to resolve overlapping contexts, finding them to be more common than orginally thought.
COMMUNITY ENGAGEMENTS
We discussed our work with participants at the "Privacy enhancing technologies in the public interest" workshop held on October 21 at Boston University, we discussed our work on both the role of contextual integrity in privacy risk assessments and the interplay between contextual integrity and differential privacy. (https://pets-for-the-public-interest-a23co.ondigitalocean.app/)
We presented our work on contextual integrity and differential privacy during the lightning talks session at the 2nd Symposium on Computer Science and Law (CSLaw) held on November 1-2 in Washington, DC. (https://computersciencelaw.org/2022/program/)
"Why the Freedom to Obfuscate is a Precious Safeguard in Digital Societies," Colloquium on Data, Ethics, and Society, College of William and Mary, October, 2022
"Contextual Integrity as a Guide to Ethical Regulation of Privacy and Data," Invited Lecture, Peking University, Law School, October 2022.
EDUCATIONAL ADVANCES
In a Ph.D. Seminar on Ethical Perspectives on Digital Technologies, students learned about the theory of contextual integrity. Three students in the class wrote research papers on applications of CI to (1) Emergency rescue drones for natural disaster situations; (2) evaluating privacy in assistive technologies for visually impaired users; (3) evaluating privacy in recording devices used in home health situations.