Towards decentralized and privacy-preserving data marketplaces to unlock data for AI: An examination of Ocean Protocol

Identifier:  TFM:952

Handle: https://hdl.handle.net/20.500.11797/TFM952

Authors:  Schwill, Frederic Christopher

Abstract:
The thesis aims to examine Ocean Protocol, an open-source toolset for privacy-preserving data exchange, especially concerning its security and privacy properties and resulting risks. First, we describe the state-of-the-art data marketplaces, their limitations and how Ocean Protocol aims to overcome them. We explain its principles and the underlying technology and show, how Ocean Protocol enables privacy-preserving data exchange. The contribution includes the local deployment of a permissionless data marketplace prototype on Ocean Protocol, which showcases a solution enabling global privacy-preserving data sharing across organizations to be used to train machine learning models. Moreover, the thesis assesses, if and under which circumstances Ocean Protocol can be used in practice to build open and privacy-preserving data marketplaces on top. The analysis primarily evaluates the resulting risks from a business perspective, with respect to corporate risk management and including relevant privacy regulation. Finally, it elaborates on how to mitigate the risks to an acceptable level and advises on how to securely share and exchange data on data marketplaces built on Ocean Protocol.