How do we know who or what to trust online & onchain?
Trust isn’t a simple yes or no thing, it’s not binary and you certainly can’t just create a universal “global trust score”. Trust is deeply personal, it’s about context and it changes depending on who you are and what you’re engaging with.
This article talks about the concept of relative trust networks and how digital attestations are changing our approach to establishing trust online.
The Concept of Relative Trust
"Relative trust is the varying degree of confidence we assign to people and things based on our personal connections & experiences.”
Think about your inner circle of friends or small community group.
When someone new comes into the picture, you start assessing: “Do we know similar people? What are their interests? Can I trust them?”. This is relative trust in its purest form, where we rely on our network of connections to navigate judgments and decisions. How much you trust can even be swayed by subtle personal cues like their energy or even how they presented themselves, whether we’re conscious of it or not. Just because you trust someone’s ability at work doesn’t mean you trust their cooking skills or that you trust them overall.
Trust is deeply personal and subjective, shaped by our unique experiences, preferences, and networks.
Trust networks aren’t just about people.
This nuanced view of trust isn't limited to personal relationships. Deciding to try a new restaurant, for instance, can be heavily influenced by friend recommendations or online reviews. This, too, is a form of relative trust, where we draw on the experiences of others to make our choices. Relative trust can also be extended to the networks of information we trust, businesses we buy from, and the applications we use.
Why a “Global Trust Score” falls flat.
We often seek guidance from "oracles of trust"—like credit agencies, data verifiers, or social media verifications/reviews—to make choices easier. But these systems, while helpful, are flawed and biased.
Take lending as an example: a bank will judge you on your credit score, but your family might lend based on personal trust, ignoring that score. And that blue check on Twitter? It doesn't mean someone's trustworthy and it’s certainly a lazy way to verify someone is not a bot.
We need context, not a one-size-fits-all score. Trust is personal and situational, and what we need is a system that understands that and gives us control.
Attestations are a key primitive for a trust network.
Attestations are simply digital signatures on structured information. They are made by one entity about someone or something. The power of this is that it’s highly composable and interoperable. You can start to compose them like Lego blocks to build context. The real value is you have the authenticity of the digital signature and the provenance of a timestamp. If you’re not familiar with attestations, you can learn more about them in our docs.
Example Attestations for a Relative Trust Network
-
A social graph and network of trusted entities
-
Verifications and approvals of a piece of information, like an article or fact.
-
Disputes and weights for not trusting certain entities or things
-
Reviews made on a particular entity
-
Trade information, interactions, and transactions between different entities
Building a Relative Trust Network
Let’s explore a practical example of how a relative trust network might form by building a social graph of people and things you trust.
Seeding the trust graph
Building a trust network starts with identifying and weighing your connections. It's like setting up your social network from scratch. You connect to a new platform designed to map out your trust relationships. It suggests entities for you to consider based on your past interactions and the digital footprints you've left — like transaction history and previous attestations.
Here's where you make it your own. You see a list of entities — people, organizations, smart contracts — and you assign a trust weight to each. It's a bit like following someone on social media but with an extra step: you're not just following; you're rating how much you trust them. Your mom might be a full 100%, a friend 85%, and vitalik.eth, whom you respect but don't know personally but respect might get a 90%.
Expanding Your Network
With these weights in place, the platform shows you a network map. It's a visual representation of your connections and their connections — a second-degree trust circle, if you will. If your friend trusts someone 80%, and you've rated your friend at 75%, the platform might suggest a 60% trust level for this new person in your network.
We even built a super simple demo of this in the early eas days.
Refining Your Trust Settings
But life isn't static, and neither is trust. You come across new entities and information that might change your perspective. Your trust network should allow you to adjust the weights dynamically. If an entity receives an attestation from a highly trusted connection, you might bump up their score. Conversely, if there's a breach of trust, you can dial it back down.
Bridging Web2 and Web3 Identities
A significant aspect of building these trust networks is bridging established Web2 reputations into the Web3 framework. EAS can help facilitate this by allowing users to bring their verified Web2 credentials into the decentralized world. We’ve seen this already with fitness data being attested to by the Receipts team and much more in development.
The Journey Has Begun, We Have the Tools.
The journey to robust trust networks begins with creating valuable attestations. Each one, especially those from trusted sources, adds a layer of depth and reliability to the network. It's not just about collecting attestations; it's the context and the credibility behind them that matter. These are the seeds from which trust networks and knowledge marketplaces sprout, offering verified content and clear information lineage.
As we build this network, privacy remains paramount. We already have the tools to ensure greater privacy and control for users. Tools like offchain attestations, the ability to generate zero-knowledge proofs from attestation data, and even using private data attestations that allow selective disclosure of data are all great approaches.
Imagine relative trust graphs of the future.
In the not-too-distant future, imagine putting on a VR headset like the Vision Pro or Quest and stepping into your trust network. This virtual space transforms abstract connections into a navigable landscape. Paths link you to individuals and entities, their brightness reflecting the strength of your trust. Intersections where paths converge mark shared connections, offering insights at a glance.
With a gesture, you could pull up a history of interactions, view attestations, and understand the context behind each connection. This VR experience would make assessing trust not just insightful but also engaging, turning the intangible into something almost tangible.
This vision is the practical evolution of trust graphs, where the digital and the experiential merge. It's a step towards making the concept of trust more accessible and interactive, allowing us to engage with our network in a way that's both meaningful and grounded in our digital reality.