TAPESTRY: Trust, Authentication and Privacy over a DeCentralised Social Registry

Project: Research

Project Details


The aim of TAPESTRY is to investigate, develop and demonstrate transformational new technologies to enable people, businesses and digital services to connect safely online, exploiting the complex "tapestry" of multi-modal signals woven by their everyday digital interactions; their Digital Personhood. In this way we will de-risk the Digital Economy, delivering completely new ways of determining or engendering trust online, and enabling users and businesses to make better decisions about who they trust online.

Online fraud and scams cost the UK economy £670m each year; crimes often perpetrated through false identities. It is difficult to make good decisions about who to trust when the digital identities of people and services are presented through pseudonyms or addresses. How can we trust that the identity we are interacting with today wasn't created out of thin air yesterday to pull a scam? Or whether the service we are registering our personal data with is trustworthy? In an era of users curating multiple digital identities that evolve over their physical lifespan, and the coming ability to migrate identities between providers (most of whom reside outside the EU), there is an urgent need for decentralised technologies to enable proofs of trust between people and services wishing to interact safely within the Digital Economy.

TAPESTRY will co-create and evaluate prototype services with end-users to determine how online behaviour and attitudes to trust could evolve in the presence of a trusted decentralised technology to prove the veracity of online identities. TAPESTRY proposes to collect, on an opt-in basis, digital trails of users' interactions (photos shared, comments left, posts 'liked', IoT devices interacted with) as encrypted trust evidence within a decentralised database (blockchain). Users grant third parties access to trust evidence for a given time period and at a given granularity, in order to prove trustworthiness of their identity via their digita personhood. For example, a crowd-funder might invite new backers to submit 2 years' history of regular social media interactions to guard against fraudulent pledging from transient identities. Community forums are becoming increasingly important for emotional support and well-being. A similar check could safeguard against trolling, or an identity posting advice could collect positive ratings within their blockchain, enabling vetting of their reputation. Deviations from behavioural norms could also be detectable within TAPESTRY to alert users to their digital identity being hacked.

From a technological standpoint, the project will develop the decentralised infrastructure necessary to make sense of the vast number of digital interactions using multimodal signals aggregated via machine learning from social media and IoT interactions. Additionally, new cryptographic strategies will be needed to secure the privacy of trust evidence and to disseminate access on a granular basis. From a HCI and co-design perspective, the development of trust services and the shift to use of the digital personhood and interaction history as trust evidence will break new ground, fundamentally altering the way users think about identity and interaction online.

To undertake this adventurous and ambitious project we have formed a strategic multi-disciplinary partnership uniting world-leading groups in multi-modal signal processing and machine learning (CVSSP), a BIS/GCHQ recognised centre of excellence for Cyber Security (SCCS), the UK's first and only 5G test-bed for next-gen mobile and IoT (ICS/5GIC), and reflecting the importance of co-designing and evaluating technology in tight integration with end-users, two leading UK groups for socio-digital interaction (DJCAD) and interaction design (UNN).

End-user partners participating in the co-design and evaluation of TAPESTRY span the technology, legal, social reform, health and well-being and commercial sectors.
Effective start/end date1/01/1730/09/20


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.