LONDON, UNITED KINGDOM – Social validation for identity is becoming increasingly popular due to its game theoretical security properties. A concept gaining traction called “Soul Bound” tokens expands the concept of digital identity significantly, and reveals the dangers of interoperable digital identity.
In a May 2022 paper titled Decentralized Society: Finding Web3’s Soul Vitalik Buterin, Glen Weyl and Puja Olhaver describe the potential of “Soulbound tokens”. Their significance as plumbing for the crypto hive minds proposed by Melianie Swan, sinks in as we investigate the details.
In this paper, we illustrate how non-transferable “soulbound” tokens (SBTs) representing the commitments, credentials, and aliations of “Souls” can encode the trust networks of the real economy to establish provenance and reputation.
Potential examples of Soulbound tokens are verifiable credentials such as learning badges, teaching certificates, employment history, works of art, photography, writings, ect. The key feature is that they are non transferable and are issued or revoked by an entity. For learning credentials this entity would be a trusted education institution, but for personal writings or works of art you would issue these yourself.
By having these tokenized credentials you build trust and reputation.
More importantly, SBTs enable other applications of increasing ambition, such as community wallet recovery, sybil-resistant governance, mechanisms for decentralization, and novel markets with decomposable, shared rights. We call this richer, pluralistic ecosystem “Decentralized Society” (DeSoc)—a co-determined sociality, where Souls and communities come together bottom-up, as emergent properties of each other to co-create plural network goods and intelligences, at a range of scales.
Some of the phraseology sounds meaningless, but will become more clear when we explore the applications of Machine Learning in this environment of soul bound and community linked identity.

The phrase “decentralized society” itself is highly weaponized, but in the technology space decentralization means building robustness into a network. A centralized network is vulnerable to attack at single points of failure, decentralized networks offer resilience through complexity. If you are designing networks for global domination you cannot risk collapse due to one building or satellite blowing up.
Even if the power or security of the network is highly decentralized, the mere creation of a standard universal network serves the same purpose as having centralized control. In other words, creating a universal standardized system means centralizing thought into one controlled and known set of patterns. Understanding this is key in order to explain why standardized technology networks violate Free Will.
Socializing Identity for Machine Computation
Many projects, such as Celo’s UBI project, Impact Markets, implement “Proof of Personhood” protocols to prevent people from creating multiple blockchain identities to take advantage of a service like UBI.
Proof of Personhood protocols (PoP) aim to provide tokens of individual uniqueness, to prevent Sybil attacks and allow non-nancialized applications. To do so, they rely on approaches such as global analysis of social graphs, biometrics, simultaneous global key parties, or some combination thereof.
However, because PoP protocols seek to represent individual identities—-focused on achieving global uniqueness—rather than social identities mapping relationships and solidarities, PoP protocols are limited to applications that treat all humans the same. Most applications we are interested in—such as staking reputation—are relational and move beyond being a unique human to being a differentiated human.
Since Soulbound tokens build a person’s trust and reputation scores through education, employment, religion and any other type of association, they expect it to become the foundation for the social credit system. Instead of just identifying an individual person they are able to differentiate people based on these relationships, which is essential for the debt markets and social impact finance.
Imagine a world where most participants have Souls that store SBTs corresponding to a series of aliations, memberships, and credentials. For example, a person might have a Soul that stores SBTs representing educational credentials, employment history, or hashes of their writings or works of art…
An ecosystem of SBTs could unlock a censorship-resistant, bottom-up alternative to top-down commercial and “social” credit systems. SBTs that represent education credentials, work history, and rental contracts could serve as a persistent record of credit-relevant history, allowing Souls to stake meaningful reputation to avoid collateral requirements and secure a loan.
The loans themselves will also constitute soulbound tokens, recording your debt arrangements in perpetuity.
Loans and credit lines could be represented as non-transferable but revocable SBTs, so they are nested amongst a Soul’s other SBTs—a kind of non-seizable reputational collateral—until they are repaid and subsequently burned, or better yet, replaced with proof of repayment.
Using Soulbound tokens for social credit profiles, as well as representing active loans, increases machine visibility to the webs of human relationships, enabling more robust algorithms calculating risk.
The ease of computing public liabilities with SBTs would open-source lending markets. New correlations between SBTs and repayment risk would emerge, birthing better lending algorithms that predict creditworthiness and thereby reduce the role of centralized, opaque credit-scoring infrastructure.
Instead of just entangling people within the thrall of global capital, they want people to incur debt relationships within their own community.
Better yet, lending would likely occur within social connections. In particular, SBTs would oer a substrate for community lending practices similar to those pioneered by Muhammad Yunus and the Grameen Bank, where members of a social network agree to support one another’s liabilities.

In addition, they want to use all these different soul bound tokens as a private key recovery system. One of the bigger problems with digital identity in general is how to “login” and how to recover your identity if you lose your password. In the blockchain universe your password is called a private key.
They point out issues with social recovery of private keys and propose an alternative where our memberships in organizations and communities can facilitate identity recovery.
A more robust solution is to tie Soul recovery to a Soul’s memberships across communities, not curating but instead drawing on a maximally broad set of real-time relationships for security. Recall that SBTs represent memberships to different communities. Some of these communities—like employers, clubs, colleges, or churches—might be more o-chain in nature, while others—like participation in protocol governance or a DAO—might be more on-chain. In a community recovery model, recovering a Soul’s private keys would require a member from a qualified majority of a (random subset of) Soul’s communities to consent.
Rights to our “souls” are governed by a subset of accrued relationships over time. High profile institutions with robust security would likely play an outsized role in identity recovery.
Moving from Artificial to Plural Intelligence
The implication of codifying social relationships is the merger of AI computation with social prediction markets. They first note how these two paradigms are limited in opposite ways. AI does not take into account economic incentives and predication markets ignore computation models. They expand:
An example of plural network goods that are of increasing salience in a digital world are predictive models built on user data. Both artificial intelligence (AI) and prediction markets seek to predict future events based on data primarily elicited from people. But both paradigms are limited in different and nearly opposite ways. The dominant paradigm in AI eschews incentives, instead hoovering up (public or privately surveilled) data feeds and synthesizing them into predictions through proprietary large-scale, non-linear models—harnessing the default web2 monopoly on “usus” without any “fructus” owing to data laborers. Prediction markets take the opposite approach, where people bet on outcome in the hopes of financial gains, relying entirely on economic incentives of financial speculation (“fructus”) without synthesizing the beliefs of bettors to produce composable models
A more productive paradigm is to eschew these extremes, and instead draw on the virtues of both, while compensating for their weaknesses and enriching their breadth. We propose thoughtfully combining the complexity of non-linear AI models with the market incentives of prediction markets to transform passive data laborers into active data creators. With such provenance-rich information rooted in the sociality of data creators, we illustrate how DeSoc can unlock plural network(ed) intelligence more powerful than either approach.
This model of combining prediction markets with AI models for use in blockchained social impact finance is already live. Alphabonds are currently in pilot with projects associated with UBS and Blackrock. The existence of soulbound tokens representing the various types of social relationships between people helps feed the AI informing impact investors how to bet on the success of a certain outcome.

Research suggests that while prediction markets generally outperform simple polling, they don’t outperform sophisticated team prediction polling, where people have incentives to share and discuss information. Under team deliberation models, members can be weighted based on factors like past performance and peer evaluation, and the team participates in semi-structured discussions to pool information that can’t be encapsulated simply in a buy or sell contract.
Whereas prediction markets elicit one number—the price of a contract—quadratic polling elicits each participant’s exact belief about the probability of an event. SBTs enable further computation over those beliefs in social context of the education credentials, memberships, and general sociality of a participant to develop better weighted (or non-linearly synthesized) predictive models, likely surfacing expert predictors at novel, unforeseen intersections
This is the basis for “Plural Intelligence” converging social prediction with artificial intelligence. They argue that in the current system, AI models view content creators as separate from their social context. Soul bound tokens show the algorithms the social context of data creators, while also enabling “governance rights” to how the collective social data is monetized. AN illusion of control and some of the profit is part of convincing the public to engage in the system, teaching the machines.
In reality almost no data is solely about individuals, all data corresponds to other people and groups as well. For example, your DNA gives information about your family, not just yourself. Soul bound tokens are a way of representing the social identity of data.
Most surveilled data creators aren’t aware of their role in creating these models, retain no residual rights, and are viewed as “incidental” rather than as key participants. Moreover, data hoovering divorces models from their social context, which masks their biases and limitations and undermines our ability to compensate for them. These tensions have increasingly come to the fore with growing demand for data availability, new initiatives like “data sheets for data sets” that document data provenance, and privacy-preserving approaches to machine learning.
Data provenance refers to information about the data such as how it was collected, where it was collected, any changes and sorts of meta-data. Soul bound tokens help represent and standardize the metadata, create economic incentives and governance rights.
Such approaches require giving meaningful economic and governance stakes to those who generate the data and incentivizing them to cooperate in producing models more powerful than what they could build alone. SBTs offer a natural way to program economic incentives for provenance-rich data while at the same time, model-makers can track the characteristics of the collected data and their social context—as reflected by SBTs—and contributors that offset biases and compensate for limits.
These economic and governance rights are called data “cooperatives”. They want as much human interaction with machine intelligence as possible. Computing different perspectives from the same set of data is important in the ambition of making AI “think” more like humans. They call it “Plural Intelligence”.
While most people do not think AI and computer networks can gain their own soul, if they are tightly integrated with Minds connected to Spirit, the computers can certainly access aspects of that incredible potential.
SBTs can also program bespoke governance rights to data creators, allowing them to form cooperatives that pool data and negotiate uses. This bottom-up programmability by data creators enables a future of plural intelligences, where model-makers can compete to negotiate uses over the same data to build different models. Thus, we move away from a paradigm of a detached monolithic “artificial intelligence” free from human origins, hoovering up provenance-free surveilled data to instead a Cambrian explosion of cooperatively constructed plural intelligences rooted in social provenance and governed by Souls.
Human and machine individuality is fuzzy as the metaverse expands. When trying to understand these designs, distinguishing between people and machines isn’t very helpful. These “plural intelligence” models eventually accrue their own soul bound tokens, building their own relationships “embedded in human sociality”. Clearly these are transhuman designs.
Over time, just as SBTs individuate a Soul, they also come to individuate models—embedding data provenance, governance and economic rights directly into the model’s code. Thus, plural intelligences—like humans—build a Soul embedded in human sociality. Or depending on how you look at it, humans evolve over time embedded in plural intelligences—each with a unique Soul, complementing and cooperating with other Souls.
If the transhuman ideology was not clear enough here’s another passage citing the founder of the ARPANET.
Through composing networks and coordination, DeSoc emerges at the intersection of politics and markets—augmenting both with sociality. DeSoc empowers the vision of JCR Licklider—founder of ARPANET that created the internet—of “man-computer symbiosis” in an “intergalactic computer network” with dramatically increased social dynamism built on trust.