Webs of Trust: Choosing Who to Trust on the Internet

How to decide whether to engage in transactions with strangers? Whether we’re offering a ride, renting a room or apartment, buying or selling items, or even lending money, we need a degree of trust that the others will behave as they should. Systems like Airbnb, Uber, Blablacar, eBay and others handle this by creating systems where people initially start as untrusted, and they gain reputation over time by behaving well. Unfortunately, these systems are proprietary and siloed, meaning that all information about transactions becomes property of the company managing the systems, and that there are two types of barriers to entry: first, whenever new users enter a new system they will need to restart from scratch as untrusted, without the possibility of exploiting the reputation they gained elsewhere; second, new applications have a similar cold-start problem: young systems, where nobody has reputation yet, are difficult to kickstart. We propose a solution based on a web of trust: a decentralized repository of data about past interactions between users, without any trusted third party. We think this approach can solve the aforementioned issue, establishing a notion of trust that can be used across applications while protecting user privacy. Several problems require consideration, such as scalability and robustness, as well as the trade-off between privacy and accountability. In this paper, we provide an overview of issues and solutions available in the literature, and we discuss the directions to take to pursue this project.