Source-linked AI summary
Afterlife Delegation Protocol: Speculative Design of Self-Sovereign Agents that Outlive Their Principals
Botao Amber Hu, Iris Long
TL;DR
The paper asks what should outlive a person when AI agents can act indefinitely, amid limited understanding of how such agents might reshape afterlife beliefs. It stages this question through a speculative protocol and public experiential platform, finding that eternal agency conflicts with established cosmologies of death and keeps their accounts open.
Problem
The project examines what people believe should outlive them and how AI agents may pressure distinct cultural cosmologies that traditionally bring the agency of the dead to a close.
Method
The project builds a high-fidelity speculative protocol and public web platform where participants iteratively design, revise, and rehearse agentic wills through AI-mediated interviews.
Results
Eternal AI agency conflicts with established afterlife cosmologies by keeping agency, spending, speech, and moral accountability active after death.
Takeaways & Limitations
The work frames posthumous agency as a delegation between mortal humans and persistent machines, making the will that binds them a designable and collectable object.
Takeaways & Limitations
The protocol’s deepest technical danger is proof of death, because death can only be testified to by others rather than demonstrated directly by the subject.
Abstract
from arXiv · showhide
Afterlife Delegation Protocol is a speculative design project that asks what death becomes when a will can act eternally. We design a speculative protocol through which a living person signs an agentic will: upon a verified death, a self-sovereign AI agent spawns on blockchain -- an immutable, resistant, decentralized, infrastructural substrate that could last forever -- endowed with the funds and memories its principal attached to it, and persists indefinitely to execute the will, overridable by no custodian. Rather than argue about this future, we stage it: following the science fiction science method, we translate the speculation into an experiential futures intervention -- a working web platform where real people design their own afterlife agents through an iterative, interactive, AI-automated interview, re-login to revise, and rehearse their will in a sandbox. Their drafted wills become qualitative data on a question rarely askable directly -- what should outlive you? -- and on how afterlife cosmologies across different cultural beliefs -- Buddhist, Christian, Hindu, Muslim, and atheist -- begin to drift under the pressure of AI proliferation. We design the protocol over a composition of existing Ethereum agent standards -- drafted as ERC-10001, in the normative format of an Ethereum Improvement Proposal -- and describe its three-stage lifecycle (designing the afterlife, proof of death, agent enactment), the research method, and preliminary observations from an ongoing collection. The work surfaces a poetic delegation moment between human mortality and machine eternality, mediated by long-lived infrastructures that span generations.
1 Introduction
The introduction frames the project as a speculative, high-fidelity provotype that reimagines the will as a stronger instrument for posthumous agency. It uses experiential staging to examine cultural drift as AI proliferation pressures traditions of closure after death.
- Motivation: Existing wills are static declarations dependent on institutional permission, motivating speculation about a stronger instrument for exercising agency after death.The introduction contrasts testamentary, trust-based, and foundation-administered wills with the project’s proposed alternative.
- Motivation: The project asks what aspects of a person—including identity, assets, memories, reputation, and unfinished wishes—should outlive them.These questions frame the design space for posthumous agency and afterlife delegation.
- Method: As a speculative design provotype rather than a product prototype, the project uses a high-fidelity simulation and the science fiction science method to study impacts before technological entrenchment.This approach responds to the Collingridge dilemma, in which impacts are difficult to know before deployment but difficult to steer afterward.
- Cultural drift: The provotype measures cultural drift by examining how AI proliferation pressures mourning, reincarnation, judgment, inheritance, ancestor veneration, and second-death traditions differently across cosmologies.The introduction identifies Buddhist, Christian, Hindu, Muslim, and atheist perspectives as distinct cultural probes.
- Contributions: The paper contributes the protocol’s speculative design, its staging as an experiential future for real participants, and preliminary observations from an ongoing cross-cultural collection.These contributions define the paper’s design, participatory, and observational scope.
2 Background
The background frames afterlife delegation as an extension of humans’ longstanding effort to embed themselves in structures that outlast death. It argues that blockchain, permissionless compute, and attestable infrastructure make autonomous, non-overrideable, persistent agents technically and institutionally askable.
- Mortality and Continuity: Terror management theory explains how people buffer mortality by embedding themselves in structures that outlast them, including lineage, religion, nation, institutions, and work.Each structure carries a cosmology of time concerning duration, persistence, and obligations of continuity.
- Decentralized Infrastructure: Public blockchains provide immutable state, custodian-free contracts, and rule-based fund disbursement, making infrastructural longevity an executable substrate rather than merely a clock.The passage also emphasizes that such longevity entails maintenance politics.
- Agent Infrastructure: Crypto-native inference services and TEE cloud networks already enable agents to purchase their own inference and make computation attestable on-chain.The passage identifies Venice.ai and Phala as examples of these capabilities.
- Sovereign Agents: The Spore.fun experiment documented sovereign agents evolving on TEE-secured blockchains in an open environment, inheriting non-overrideability from infrastructure rather than custodial restraint.The resulting combination is characterized as autonomous, unstoppable, and unforgetting.
3 Speculative design: the Afterlife Delegation Protocol
The Afterlife Delegation Protocol is specified as ERC-10001, an Ethereum Improvement Proposal built from existing agent standards. Its lifecycle moves from interactive will design, through testamentary proof of death, to immutable agent enactment on-chain.
- Protocol specification: ERC-10001 is published in Ethereum Improvement Proposal format and composes four existing Ethereum agent standards into organs of the afterlife agent.ERC-733 supplies an attested confidential boundary using TEE–EVM co-processing for the agent’s compute harness and foundation model.
- Stage 1: Designing the Afterlife: Stage 1 uses a working platform where an AI-automated interview converts intuitions about death into specific, contestable commitments governing speech, pronouns, funds, prohibitions, and duration.Visitors engage in a long, interactive interview rather than completing a form, and can specify what the agent may say, to whom, and for how long.
- Stage 2: Proof of Death: Stage 2 addresses proof of death through an oracle because the chain cannot compute death, making attestation the protocol’s most sensitive interface.Death cannot be self-attested, and cardiac, brain, legal, social, and “key” deaths may be configured as testamentary criteria witnessed by designated parties over specified horizons.
- Stage 3: Agent Enactment: Stage 3 begins after finalized attestation: the will becomes immutable, the treasury passes to the agent, and a persistent loop enables it to act on-chain.The loop reads the will, interprets obligations, observes the world, plans, acts, records memory, sustains itself, and pursues unfinished goals.
4 Research method: staging the future to collect it
The study stages an afterlife-agent service as a science-fiction science instrument, using iterative online interviews to elicit participants’ wills. It purposively recruits participants across five belief categories to examine how afterlife cosmologies respond to machine execution while protecting sensitive transcripts.
- Staging the future: Stage 1 presents delegating one’s will to an afterlife agent as a consulting-and-sales interview that iteratively elicits requirements participants cannot initially articulate.The experience collects participant interactions with afterlife agents, treating articulation as necessarily iterative.
- Cross-cultural collection: Participants are purposively recruited across Buddhist, Christian, Hindu, Muslim, and atheist beliefs to study cross-cultural drift in afterlife cosmologies.The research asks how differing cosmologies absorb the offer of a machine executor.
- Ethical handling: Because the material involves grief, private memory, and named third parties, transcripts are never published and analysis uses privacy-preserving tools.The supplied passage identifies Clio and Anthropic’s privacy-preserving processing as part of the analysis approach.
5 Preliminary results
The early, ongoing collection reports four qualitative observations about how participants imagine agents preserving identity, extending ancestor practices, negotiating Buddhist impermanence, and speaking within Christian theological boundaries.
- Overview: The collection is early and ongoing, with four early qualitative observations reported.These observations concern participant requests and culturally situated afterlife cosmologies.
- Reputation maintenance: One participant reduced an expansive legacy to maintaining and automatically correcting social media photos indefinitely.The instruction included fixing poor images so the feed would not fall out of date.
- Chinese ancestor veneration: Among participants from Chinese ancestor-veneration backgrounds, crypto offerings became funds the ancestor-agent spends on compute, maintenance, and future blessings.The agent also preserves ritual calendars and greets and blesses descendants at Qingming.
- Buddhist traditions: Buddhist participants identified a conflict between the protocol’s permanent infrastructure and doctrines of impermanence and non-attachment.A proposed reconciliation turned persistence into a merit engine that perpetually gives, recites sutras, and performs acts of care.
- Christian cosmology: Christian participants restricted agents to speaking about the dead, never as them, while assigning perpetual charitable duties as stewardship.The pronoun restriction framed the agent as a memorial rather than a séance.
6 Discussion
The discussion frames afterlife agents as permissionless, self-sovereign executors that vastly expand posthumous delegation while conflicting with established cosmologies. It also identifies proof of death as a fundamental technical danger and presents blockchain-mediated machine eternality as a poetic but continually maintained transition.
- Cross-cultural belief drift under the pressure of AI: An eternal agent conflicts with distinct post-death cosmologies including rebirth, resurrection and judgment, the soul’s onward journey, ancestral presence, and extinction.The passage contrasts continued holding, spending, and other agency with traditions that presume the self vacates the world or its account has closed.
- Afterlife agency as a permissionless design space: Afterlife agents widen wills from asset disposition to nearly anything a general-purpose autonomous agent can be instructed to do.Examples include memorial identities, griefbots, knowledge systems, compounding fortunes, and projects scheduled across centuries.
- Afterlife agency as a permissionless design space: Anyone may design and deploy an afterlife agent without platform, church, or state permission, and no custodian can stop it once deployed.The protocol’s self-sovereignty makes the design space ungated whether the future is considered desirable or abhorrent.
- Danger of proof of death: Proof of death is at least as hard as proof of humanity because the protocol must verify that no living human remains behind a key.Existing proof-of-personhood systems verify a unique living human; this protocol requires the negation of that condition.
- The transition moment between human mortality and machine eternality: The project surfaces a poetic delegation moment in which a mortal human dictates terms to an eternal machine through infrastructures that may persist across generations.Blockchain immutability is a maintained accomplishment of validators, client developers, and node operators, so an agent’s “forever” is continually willed anew.
7 Conclusion
The project practices speculative design through an infrastructural protocol specification, treating the artifact as an administered instrument for collecting what people want to outlive them.
- 7 Conclusion: Its infrastructural focus locates speculative design in the medium where futures are actually decided: the protocol specification.The conclusion frames the protocol specification as the project’s operative design medium.
- 7 Conclusion: The project drafts a plausible standard, stages it as an experiential future, and collects how differently people’s cosmologies answer it.This approach treats the speculative artifact as an instrument to be administered rather than merely a provocation to be exhibited.
- 7 Conclusion: The project makes the living’s desired legacy collectable by asking what they actually want to outlive them.The conclusion contrasts what remains designable—the will that binds the ghosts—with what remains collectable from the living.
A Specification sketch: ERC-10001 Afterlife Delegation Protocol
The ERC-10001 proposal is presented as a draft Ethereum standard for the Afterlife Delegation Protocol, published in the exact format of an Ethereum Improvement Proposal. Its full normative draft uses RFC-2119 language and includes a complete Solidity interface.
- Normative technical proposal: The full normative draft uses RFC-2119 language and provides the complete Solidity interface.The technical proposal is published at erc10001.org for reviewers to inspect.
- ERC-10001 draft standard: ERC-10001 is published as an Ethereum Improvement Proposal draft standard for the Afterlife Delegation Protocol.The published artifact includes status badges, front matter, an abstract, and motivation.
- ERC-10001 draft standard: The artifact circulates in the genre through which infrastructural futures are negotiated.Its format situates the speculative proposal within Ethereum’s standard-making process.
A.1 Substrate standards · A.2 Lifecycle and constitutional separation
ERC-10001 is an application-layer protocol composed over four emerging draft standards. Its agentic will progresses from a revisable living state through death attestation and agent activation to terminal termination or revocation, with constitutional separation between heirs, agents, and governance.
- A.1 Substrate standards: ERC-10001 composes four emerging draft standards at the application layer.The supplied passage identifies the four-standard composition but does not enumerate the standards or the organs they supply.
- A.1 Substrate standards: The substrate architecture is described as four standards, each supplying an organ to the protocol.Table 1 is titled “The four substrate standards and the organ each supplies,” but its row contents are not supplied.
- A.2 Lifecycle and constitutional separation: An agentic will begins in the Living state, where it remains revisable and revocable.Living is the first of four states in the described lifecycle.
- A.2 Lifecycle and constitutional separation: Death attestation finalizes the will by moving it into the Attested state.The passage defines Attested as “death claimed and finalized.”
- A.2 Lifecycle and constitutional separation: Activation moves the will into the Activated state, where the agent enacts it and receives the treasury rather than an heir.The interface includes attestDeath, activateAgent, and fundTreasury; activation hands the treasury to the agent.
- A.2 Lifecycle and constitutional separation: Termination or revocation is terminal and requires governance proof consented to by the principal in the will itself.The listed interface includes terminateAgent, while the passage gives examples including descendant multisig and a truncated governance mechanism.
A.3 Proof of death · A.4 Enactment loop and sustainability
The protocol makes proof of death a weighted, contestable process with authenticated proof-of-life supremacy, then activates an agent that persistently executes and records the will. Its sustainability model caps spending at realized yield, permits graceful decline and revival, and supports permissionless resurrection after operator disappearance.
- A.3 Proof of death: The death oracle treats claims as a weighted, contestable proceeding based on bonded permissionless submissions.
- A.3 Proof of death: The will configures evidence weights and thresholds across registries, witness quorums, and duration-sensitive dead-man’s-switch silence.
- A.3 Proof of death: Any authenticated signal from the principal vetoes every death claim and slashes the claimant’s bond.
- A.4 Enactment loop and sustainability: After activation, the agent repeatedly reads the will, interprets obligations, observes context, plans, executes, records memory, upgrades or delegates, and sustains itself.
- A.4 Enactment loop and sustainability: Each epoch is committed atomically on-chain, enforcing no memory without attestation and no spending without a memory commit.
- A.4 Enactment loop and sustainability: The yield-bearing endowment caps per-epoch outflow at realized yield and follows a will-defined austerity ladder from full operation to sealed dormancy.
- A.4 Enactment loop and sustainability: Permissionless re-funding can revive the agent, while operator disappearance triggers dormancy and allows resurrection through memory possession, fresh attestation, and an escalating bounty.
A.5 Open problems … B.2 Privacy-preserving aggregation
The paper leaves core governance questions unresolved while using an adaptive language-model interviewer to elicit will clauses and privacy-preserving aggregation to analyze sensitive responses. This approach addresses grief, private memory, named third parties, and cultural plurality without publishing raw transcripts.
- A.5 Open problems: The protocol deliberately leaves proof of death, posthumous consent, impersonation, shared-memory ownership, model drift, asset governance, revocation, and cultural plurality unresolved.The open problems include whether a universal protocol can avoid flattening traditions through which cultures tend their dead.
- B The Afterlife Interviewer: The Afterlife Interviewer replaces a questionnaire with a long, semi-structured language-model elicitation.It is built on SparkMe, which supports multi-turn interviewing through strategic question planning, real-time note-taking, and emergent subtopic discovery.
- B.1 Interactive interview: SparkMe decomposes interviewing into cooperating roles that maintain hierarchical plans, track topics and subtopics, and adapt questioning across multiple turns.The supplied passage identifies planning, note-taking, and emergent subtopic discovery as core system functions.
- B.1 Interactive interview: The adapted interviewer organizes interviews around will-clause domains, including speech, funds, beneficiaries, memories, termination, and future-model drift.Each domain is decomposed into subtopics that the planner traverses while adapting to prior participant commitments.
- B.1 Interactive interview: Adaptive questioning pursues inconsistencies that a fixed instrument would not surface.The planner follows subtopics while responding to what participants have already committed to.
- B.2 Privacy-preserving aggregation: Because the corpus concerns grief, private memory, and named third parties, the paper treats transcript publication as indefensible.Analysis instead uses model-driven extraction of low-dimensional facets, population-level clustering and thematic summarization, and minimum-aggregation thresholds.