Reputation, Trust, and Privacy in the Sharing Economy: A Network Science Approach

This project will develop a complete statistical description of distributed online systems where interactions between users are driven by reputation. Such systems are epitomised by the emerging online marketplaces of the sharing economy, such as Airbnb or Uber, where "micro entrepreneurs" and customers build a reputation through an online peer-review process.

Recent estimates project revenues from the top five sectors of the UK sharing economy to reach £9 billion within the next ten years. Such a fast growth will dramatically increase the interconnectedness between different online marketplaces and their users. This, in turn, will bring about the need to promote trust on large scales by merging the reputations developed by the same users on different platforms. Indeed, a few startup companies already offer embryonic services whose users receive a portable reputation score based on the aggregate of their public online activity. Similar practices will require digital personhood to become more and more transparent to others, with serious implications to online privacy.

This project will address the interplay between reputation, trust, and privacy lying at the core of the sharing economy. In order to do so, it will start from the observation that the sharing economy is a large complex network of interactions. As such, it falls squarely within the realm of application of Statistical Physics and Complexity Science, where collective macroscopic behaviour emerges from local interactions between elementary components. By taking this perspective, this project will produce a network vision of the sharing economy by first analysing data from platforms where reputation-driven interactions are at play, and by studying how reputation and trust between users form in online environments with the methods of Experimental Psychology. Then, by building upon this empirical and behavioural knowledge, network models capable of reproducing and predicting the macroscopic behaviour of complex online marketplaces will be designed.

The project's high-level objectives are:

1. To identify the main empirical and behavioural regularities of online marketplaces driven by reputation

2. To model trust and efficiency in sharing economies as collective network phenomena emerging from the interactions between users

3. To model the effects of reputation aggregation in realistic multi-platform sharing economies

The main non-academic beneficiaries of the proposed research will be startup digital companies active in the sharing economy sector. At present, several players have implemented notions of reputation and trustworthiness in online marketplaces. However, most of them are individual-centric metrics that do not take into account the crucial impact of indirect interactions with others. Several modeling efforts throughout the Fellowship will lead to the identification of notions of reputation and trust that fully account for the complex network structures supporting the interactions between users of the sharing economy. Industrial impact will therefore be achieved with the uptake by relevant players of the main ideas put forward in the proposed research plan, which will be promoted through the organisation of industry-academia workshops, and the engagement with some of the major startup accelerators in the London area.

The later stages of the Fellowship will focus on reputation aggregation, a budding practice in the sharing economy which entails the merging into portable aggregate scores of the reputations developed by a single user in different contexts. The proposed research plan puts forward the first systematic study of this practice, which is currently implemented by a few digital companies via black-box algorithms. The project's results in terms of reputation aggregation will seek to outperform the existing standards by devising protocols that prevent users from being exceedingly rewarded or marginalised across multiple platforms due to good or bad performance in a fraction of them. This output will provide guidelines for practical solutions that are likely to have the potential for commercialisation, which will be sought through the launch of an academic spinoff.

Parts of the project will also provide useful input to upcoming policy-making processes. Indeed, both the UK private and public sector have recently begun addressing the need for regulation in the sharing economy and its possible scope. In partnership with the OECD Directorate for Science, Technology and Innovation, the Fellowship will model several aspects of the sharing economy that are likely to be addressed by regulators in the near future, such as privacy requirements, online trust, and the aforementioned practice of reputation aggregation. Models will be designed in order to allow for the easy numerical experimentation of different policy options.

In summary, the main beneficiaries from the output of the Fellowship are expected to be:

1. Emerging digital companies through the uptake of results and guidelines put forward by the proposed research

2. Users and customers of sharing economy platforms, through the commercialisation of the Fellowship's core results on reputation aggregation

3. Policy-makers through the dissemination of the Fellowship's ideas that are of relevance to the upcoming regulatory process of the sharing economy