I am Haojian Jin, an Assistant Professor in Halıcıoğlu Data Science Institute at UC San Diego. I recently graduated from Human-Computer Interaction Institute at Carnegie Mellon University. My Ph.D. dissertation proposes a new way to design, implement, and interact with permissions in the interest of privacy.
I hope to lead discussions on creating a virtuous cycle ecosystem for privacy. Namely, I am interested in enabling a new ecosystem where building trustworthy systems is rewarded, and developers compete to guarantee greater user protection, not less. This discussion falls under the proposed scope - permission transparency, accountability, and control.
My position statement has three key ideas.
-
First, future developers should collect users' data on a need-to-know basis and implement it in a transparent and structured manner.
-
Second, we should not expect users to manage permissions actively. Many studies find that users do not have the expertise and time. I argue that many low-risk fine-grained permissions can be granted by default. Our goal, instead, is to empower third-party auditors and privacy advocates in the ecosystem and disincentivize developers from using unnecessary permissions. For example, consumer reports may rank app developers by privacy. App stores can inform users of the potential privacy risks during installation.
-
Third, most users will rely on trusted third parties to manage their permissions aggregately. For example, a trusted third party may develop Privacy Diagnostics, similar to security diagnostics in anti-virus software, to help users identify and fix suspicious permissions.
The discussion would be more about "what if" rather than "how." Here I enumerate a few potential discussion examples:
- What are the potential drawbacks of this new paradigm?
- When would this new paradigm fail?
- What are low-risk fine-grained permissions?
- When do we need approvals from users? For example, assuming a system offers a differential privacy guarantee, can we grant permissions by default?
-
Modular Privacy Flows - 3 page summary of my Ph.D. dissertation
- This document describes "how" to enable the virtuous cycle ecosystem. The main new idea is to replace all-or-nothing binary APIs with operator-based APIs.
- Operator-based Data Access for Protecting Users' Privacy further summarizes the idea of creating a virtuous cycle ecosystem for privacy.
-
- Peekaboo is a prototype system I have built to demonstrate the vision I described above. It leverages an in-home hub to pre-process and minimize outgoing data in a structured and enforceable manner before sending it to external cloud servers.
-
- This paper compares 11 different privacy protection concepts with 227 participants, offers a list of the pros and cons for each concept, and ranks these design concepts.
I hope to have some technical discussion about different implementation strategies (e.g., P3P, IoT matter) and their tradeoffs.