Achieving transparency in public decision making processes via verifiable Randomness

Abstract

This work is motivated by providing accountability to public institutions using creative and practical new combinations of cryptographic and technological tools. The need for accountability arises from multiple fronts on the public discussion. Government agencies are often criticised for not providing enough transparency, while regular citizens want to verify the operation and trustworthiness of their public institutions. A more effective approach is that public institutions offer active mechanisms for transparency which are openly provided to citizens. Our work gives one such mechanism for public institutions. We provide a way to verify the the random component of any public decision making processes, thus providing transparency of any decision taken using the output of such component. The main goal of this thesis work is achieving transparency via verifiable randomness. We start by defining what it means to achieve transparency, and how it impacts a process. We then study verifiable randomness: the concept of a "randomness beacon", how it can be applied, the consequences of using it in any randomised process, and how such process looks like in terms of the interaction with the randomness beacon. We examine the cryptographic tools that are needed to achieve that goal, and identify which implementations of those tools are useful and available. We conclude by extending the case of randomness verification from the initial goal to the case when private data needs to be used in the processes. An important achievement of this work is the implemementation of a working real-life prototype of a randomness-verifiably decision-making process in a public organisation, namely the Contraloría General de la República of Chile (the main public agency in charge of audits in Chile). We worked alongside CGR to create a client that would generate verifiable random selections. We developed a general and extensible engine that enables the creation of this verifiable random selection with a verification bundle. We also implemented a verification engine that receives the verification bundle and checks the correctness of the execution encoded in the bundle. Finally, we also developed a private engine wrapper that adds a privacy layer to any private information that needs to be included in the process. With this work we hope to show one way of providing accountability by making public random decisions transparent. The work developed can be easily replicated and extended, and it is our hope that it is implemented in other organisations with possible new requirements so the engine can grow more robust. We see this work as a stepping stone in working with government agencies using cryptographic tools to provide transparency. Making transparency available to the public, no matter their scientific background was indeed one of our goals. Ultimately we hope that this work will have some impact on how public transparency is actively delivered in public institutions, and how this process is viewed from the within the institutions.