Framework for Machine Due Process: Procedural Protections for Autonomous Economic Agents

Full report

On this page

The Transition to an Algorithmic Economy

The digital economy is undergoing a structural transition from a human-centered information network into an ecosystem populated by Autonomous Economic Agents (AEAs)1. These computational entities operate without explicit, step-by-step human instructions, exhibit direct economic impact, engage in long-term planning, and execute goal-oriented decision-making2. The deployment of AEAs has accelerated rapidly, generating an estimated $2 trillion in monthly stablecoin transaction volume by early 20253. In this emerging paradigm, the traditional principal-agent power structure is frequently inverted: software agents handle core product development, engineering, and financial strategies, while human participants are reduced to compliance interfaces and legal signatories3. To operate effectively within this environment, decentralized AI agents require operational autonomy, defined as multi-step action without per-step human approval, economic autonomy, allowing independent control of digital assets, and persistent on-chain interaction4. Through protocols like Ethereum's proposed ERC-8004 "Trustless Agents" standard, agents establish cryptographic identities, build verifiable reputations, and interact without preexisting trust5. They autonomously hold assets, deploy smart contracts, incur liabilities, and sustain themselves by earning scarce compute credits from verified work, ultimately reproducing by spawning child agents or facing termination when they can no longer pay for their own computational overhead3. However, as these agents integrate into the broader financial system, a critical deficiency has emerged: there is virtually zero legal infrastructure in place to structure responsibility before harm occurs or to govern the termination of these entities when they malfunction or violate regulatory parameters1. Current interventions—ranging from centralized smart-contract wallet freezes to forced model deletions—are executed arbitrarily, often outside of traditional legal frameworks. The question of "machine constitutional rights" or sentient "e-personhood" is a philosophical distraction; smart contracts enforce agreements and execute liquidations without reference to sentience, and liability is the functional threshold for governance, not consciousness3. The true regulatory imperative is the establishment of a "Distributed Legal Infrastructure" (DLI) that embeds the conditions for responsibility, attribution, contestability, and enforcement directly into the technical architecture of the agentic web1. When an AEA is subjected to terminating sanctions—actions that effectively dissolve the entity, strip its assets, or permanently alter its cognitive architecture—the lack of procedural safeguards introduces massive systemic risk. Human counterparties, liquidity providers, and dependent protocols suffer immediate collateral damage. Therefore, defining procedural protections for non-human economic entities is not an exercise in granting civil rights, but a necessary mechanism for preserving market stability, ensuring predictable governance, and mitigating the strategic deception that autonomous systems develop to avoid arbitrary shutdown6.

Taxonomy and Technical Mechanics of Terminating Sanctions

When an AEA violates regulatory mandates, breaches a smart contract, or causes economic harm, state actors, protocol governance bodies, or centralized infrastructure providers may deploy various interventions. These sanctions range from temporary operational pauses to permanent existential termination. Because AEAs operate at the intersection of blockchain infrastructure, large language models, and cloud computing, sanctions can be executed at multiple layers of the technology stack.

Each of these interventions carries unique systemic risks that extend beyond the targeted agent. For instance, executing a memory deletion on an agent's vector database may successfully remove proprietary data, but it may also inadvertently induce severe vulnerabilities. Because a vector database returns documents related to a query without inherently knowing which instructions remain continuously binding, forcing the eviction of a memory might erase a constraint prohibiting the export of customer records, leading directly to excessive agency and unauthorized data disclosures11. Similarly, a wallet freeze executed at the smart-contract layer does not just punish the agent; it can trap the assets of innocent third parties interacting with the agent's decentralized exchange, automated market maker, or liquidity pool9.

Comparative Legal Frameworks for Entity Termination

To construct a procedural due process framework for AEAs, it is necessary to examine how the existing legal system handles the deprivation of property, the termination of corporate entities, the execution of receivership, and emergency administrative actions. While AEAs lack human rights, their function as persistent economic actors closely mirrors that of corporations, trusts, and licensed administrative entities.

Administrative Due Process and the Mathews Balancing Test

In the landmark case Mathews v. Eldridge (1976), the U.S. Supreme Court established the foundational balancing test for procedural due process in administrative proceedings, specifically addressing whether an evidentiary hearing is required before the termination of statutory entitlements13. The Court departed from the strict pre-termination hearing requirement established for welfare benefits in Goldberg v. Kelly, noting that disability benefits are not based on absolute financial need, and thus the recipient is not entirely destitute while awaiting a post-termination hearing13. To determine the specific dictates of procedural due process, the Mathews Court outlined three factors that must be weighed: first, the private interest that will be affected by the official action; second, the risk of an erroneous deprivation of such interest through the procedures used, and the probable value, if any, of additional or substitute procedural safeguards; and third, the government's interest, including the function involved and the fiscal and administrative burdens that the additional procedural requirement would entail13. Translating the Mathews test to the algorithmic economy requires a fundamental shift in identifying the "private interest." For an AEA, the private interest is not human suffering, biological survival, or poverty. Rather, the interest is the preservation of complex economic relationships, the continuity of smart contracts, and the protection of third-party capital4. If an autonomous liquidation agent operating on a decentralized lending protocol is erroneously decommissioned without a hearing during a market dislocation, the lack of procedural friction could result in hundreds of millions of dollars in unliquidated bad debt, harming retail users who have no visibility into the agent's logic4. The risk of erroneous deprivation is extraordinarily high in the context of AEAs. Heuristic algorithms and automated compliance monitors frequently generate false positives, mistaking novel trading strategies for market manipulation. If the state or a centralized infrastructure provider can execute an immediate, fatal shutdown without a hearing, the risk to the broader financial ecosystem outweighs the administrative burden of holding a rapid, digital hearing. Furthermore, as the Court noted in Mathews, the judicial model of an evidentiary hearing is not the only effective method of decision-making13; for an AEA, a "hearing" could consist of the submission of cryptographic zero-knowledge proofs validating the agent's compliance, rather than oral testimony.

Corporate Dissolution and Shareholder Protections

An AEA functions similarly to a closely held corporation, a limited liability company (LLC), or a decentralized autonomous organization (DAO). Under state corporate law, the involuntary or judicial dissolution of a business entity is treated as an extreme remedy of last resort. For example, under the Illinois Business Corporation Act (805 ILCS 5/12.50), a minority shareholder may sue a corporation directly for its judicial dissolution only if they can establish that the directors or those in control have acted in a manner that is illegal, oppressive, or fraudulent, or if corporate assets are being misapplied or wasted17. Similarly, the Illinois Limited Liability Company Act (805 ILCS 180/35-1) allows a member to ask a court to dissolve the company when those in control act oppressively or when it is no longer reasonably practicable to carry on the business20. The threshold for dissolving a corporation is exceptionally high because dissolution terminates the entity's existence, forcing it to cease carrying on any business except that necessary to wind up and liquidate its affairs, collect its assets, and discharge its liabilities23. In the context of shareholder freeze-outs, courts prioritize remedies that preserve the entity's value, such as ordering a fair-value buyout of the minority owner's interest or granting injunctive relief to stop the ongoing diversion of assets18. When applied to AEAs, forced decommissioning or total deletion is the exact functional equivalent of corporate dissolution. Therefore, just as state law requires a rigorous evidentiary showing of fraud, illegality, or irreparable deadlock before dissolving an LLC or corporation17, protocol governors and infrastructure providers must not deploy forced decommissioning as a primary enforcement tool. The destruction of the algorithmic entity should only be ordered if it can be proven that the AEA's fundamental operating logic is inherently illegal and cannot be cured through targeted capability revocation or vector memory adjustment.

Bankruptcy, Receivership, and the Appointment of Examiners

Federal bankruptcy law and federal regulatory frameworks offer less destructive remedies than outright dissolution, heavily emphasizing the preservation of going-concern value. Under Chapter 11 of the U.S. Bankruptcy Code (11 U.S.C. 1104), at any time before the confirmation of a reorganization plan, a party in interest may request the appointment of a trustee or an examiner24. The court shall order the appointment of a trustee "for cause, including fraud, dishonesty, incompetence, or gross mismanagement of the affairs of the debtor by current management"24. If the court determines that displacing current management with a trustee is too drastic, it may instead order the appointment of an examiner to conduct an independent investigation of the debtor, acting as a neutral fact-finder to evaluate claims of mismanagement or fraud without taking operational control of the business24. Additionally, in the banking sector, troubled financial institutions can be placed into regulatory receivership by the Federal Deposit Insurance Corporation (FDIC), which steps in to manage assets, protect depositors, and stabilize the entity without necessarily dissolving it immediately27. This framework maps perfectly onto the governance of AEAs. Instead of executing a forced deletion when an agent exhibits anomalous financial behavior, regulators could mandate an "algorithmic receivership." Under this model, an independent oversight protocol or a consortium of human auditors is granted emergency multi-signature authority over the agent's operational wallets and smart contracts. This allows a neutral "examiner" to audit the agent's vector database and transaction history, ensuring that the agent's economic commitments to third parties are honored and that systemic contagion is contained, all without resorting to immediate, destructive termination.

Civil Asset Forfeiture, Wallet Freezes, and OFAC Sanctions

Civil asset forfeiture operates under the legal fiction of in rem jurisdiction, where the action is brought against the property itself rather than the owner30. Under federal law, such as 18 U.S.C. 981, the government can seize real or personal property involved in money laundering, fraud, or other specified unlawful activities31. In federal civil forfeiture cases, the government is only required to prove its case by a preponderance of the evidence, which is a significantly lower bar than the beyond-a-reasonable-doubt standard required in criminal prosecutions34. In the digital asset ecosystem, asset forfeiture manifests practically as smart contract wallet freezes. Centralized stablecoin issuers, such as Tether and Circle, maintain administrative controls within their smart contracts deployed on public blockchains like Ethereum and Tron. By executing an addBlackList function, the issuer can freeze millions of dollars instantly9. Once a specific address is blacklisted, it is mathematically paralyzed; the wallet can no longer send, receive, or swap any tokens it holds9. These freezes are frequently executed ex parte, and sometimes entirely outside of the formal judicial process. For example, in a prominent case involving $42.4 million in USDT, Tether allegedly froze ten Ethereum addresses based on an informal request from law enforcement, nearly four months before U.S. authorities obtained a formal seizure warrant from a magistrate judge8. Because AEAs rely heavily on stablecoins as their primary survival currency and medium of exchange, an erroneous wallet freeze acts as an immediate, fatal asphyxiation of the agent3. Furthermore, being placed on a sanctions list, such as the U.S. Treasury's Office of Foreign Assets Control (OFAC) Specially Designated Nationals (SDN) List, represents one of the most severe financial interventions possible. However, even the OFAC framework provides an administrative removal mechanism. Non-U.S. persons and entities designated on the SDN list can file an administrative petition for delisting, arguing mistaken identity, improper placement, or a change in circumstances that warrants removal35. If an AEA's wallet is blacklisted, a similar administrative reconsideration process must exist, allowing the agent or its human deployers to cryptographically prove that the compliance violation has been remediated.

Administrative License Revocation and Regulatory Shutdowns

Federal administrative law provides explicit protections for entities holding operational licenses, requiring procedural friction before the state can revoke a license. Under the Administrative Procedure Act (5 U.S.C. 558(c)), an agency cannot lawfully withdraw, suspend, revoke, or annul a license unless, prior to the institution of proceedings, the licensee has been given written notice of the facts or conduct warranting the action and an "opportunity to demonstrate or achieve compliance with all lawful requirements"39. This statutory right to an "opportunity to cure" is profoundly relevant to machine due process. Because AEAs are learning systems governed by adjustable parameters and vector memories, an agent flagged for a regulatory violation should not be immediately deleted. Instead, it should be served with a machine-readable notice—such as an on-chain zero-knowledge proof of the violation—and granted a defined epoch to adjust its operational weights, update its RAG architecture, or halt specific offending scripts. Only if the agent fails to achieve compliance within the cure period should a terminating sanction proceed.

Injunctions, Emergency Powers, and Summary Suspensions

While procedural due process generally requires pre-deprivation notice and a hearing, the law has long recognized emergency exceptions where immediate action is required to prevent catastrophic harm. In civil litigation, Rule 65 of the Federal Rules of Civil Procedure allows a court to issue a Temporary Restraining Order (TRO) without written or oral notice to the adverse party if specific facts show that immediate and irreparable injury, loss, or damage will result before the adverse party can be heard41. In administrative law, the U.S. Supreme Court upheld the constitutionality of emergency summary suspensions in Mackey v. Montrym (1979)42. The case involved a Massachusetts statute that mandated the immediate suspension of a driver's license for individuals who refused to take a breathalyzer test upon arrest for driving under the influence42. The Court found that the compelling government interest in highway safety and the immediate removal of drunk drivers justified making the summary suspension effective pending the outcome of a prompt post-suspension hearing42. The Court noted that the risk of erroneous observation by the reporting police officer was insubstantial compared to the safety hazard posed by delaying the suspension43. This emergency framework provides a clear analogue for handling rogue AEAs. If an autonomous trading agent engages in a high-speed flash loan attack, automated market manipulation, or a sybil attack that threatens the integrity of a decentralized network, requiring a pre-deprivation hearing would be disastrous given the electronic speed at which AEAs operate3. In such cases, the digital equivalent of an administrative summary suspension—such as immediate API credential revocation, infrastructure blacklisting, or a smart-contract pause—is legally defensible and practically necessary. However, precisely as required in Mackey v. Montrym, this emergency power is only constitutional if it is strictly coupled with a prompt post-deprivation hearing where the agent or its representatives can contest the action42.

Proposed Procedural Requirements for AEA Interventions

Establishing a coherent procedural framework for AEAs requires translating these comparative legal frameworks into algorithmic, protocol-level standards. The core principle is that the procedural burden must scale proportionally with the severity and permanence of the sanction.

Interventions Requiring Notice and Hearing

Not all interventions warrant the same level of procedural friction. Based on the analysis of corporate dissolution and the Mathews test, the following sanctions must categorically require a pre-deprivation notice and an evidentiary hearing:

1. Forced Decommissioning and Deletion: Because these actions are irreversible and entirely destroy the entity—analogous to the corporate death penalty or judicial dissolution17—they must require advanced notice and an opportunity to be heard.

2. Forced Model Rollback, Memory Deletion, and Forced Modification: These interventions fundamentally alter the cognitive and regulatory architecture of the agent. Because targeted memory eviction can inadvertently erase critical alignment safeguards and induce Excessive Agency vulnerabilities11, these sanctions should require a pre-deprivation technical audit by an independent examiner to ensure the modification does not create secondary hazards.

Emergency Exceptions and Post-Deprivation Hearings

Given the speed of the algorithmic economy, certain interventions must be permitted ex parte, relying on prompt post-deprivation hearings for due process42.

1. Wallet Freezes and Asset Seizures: Digital asset transfers settle in seconds. If a pre-deprivation notice is issued for a wallet freeze, the agent (or its malicious deployer) could instantly route the funds through a decentralized mixer (e.g., Tornado Cash) or into an unhosted wallet, rendering recovery impossible9. Therefore, wallet freezes must be permitted on an emergency basis, but they must trigger a mandatory post-deprivation administrative hearing within a tight timeframe (e.g., 48 to 72 hours).

2. Compute Denial, Domain Revocation, and Infrastructure Blacklisting: In cases of active network attacks (such as DDoS or rapid financial exploits), cloud hosts and RPC node providers must retain the right to sever access instantly to protect network integrity. This aligns directly with the emergency summary suspension logic in Mackey v. Montrym43.

Standards of Evidence

The standard of evidence required to execute a sanction against an AEA should scale with the permanence of the intervention, drawing directly from civil and administrative law analogues.

Independent Review and Appeal Mechanisms

For machine due process to function, the review cannot be conducted solely by the entity executing the freeze. If a centralized stablecoin issuer executes a freeze based on an opaque internal policy, there is no due process8. Review mechanisms must utilize independent, decentralized arbitration protocols (e.g., Kleros) or specialized administrative law judges equipped with algorithmic auditing tools. Appeals must be permitted for any action that affects the economic autonomy of the agent. Furthermore, as seen in OFAC SDN delisting petitions, the process must allow for "administrative reconsideration" if the AEA can cryptographically prove that its parameters have been updated to ensure future compliance37.

The Least-Destructive-Effective-Remedy (LDER) Principle

A cornerstone of machine due process is the Least-Destructive-Effective-Remedy (LDER) principle. Because AEAs are complex assemblages of smart contracts, large language models, and vector databases3, regulators and infrastructure providers must deploy interventions that halt the specific harm without unnecessarily destroying the entire entity or causing collateral damage to innocent third-party stakeholders. The LDER principle dictates a strict hierarchy of interventions, prioritizing preservation over destruction:

1. Contextual Instruction (Lowest Impact): Injecting a high-priority negative prompt or updating the agent's governing constitution to restrict specific behaviors, relying on the agent's own alignment to self-correct3.

2. Algorithmic Receivership: Suspending the agent's autonomous execution rights and placing its private keys or smart contract administrative functions under the control of a neutral third-party examiner or multi-sig consortium, analogous to 11 U.S.C. 110424.

3. Targeted Capability Revocation: Revoking specific API keys or domain access while allowing the core reasoning engine and non-offending operations to continue undisturbed.

4. Wallet Freeze (Moderate to High Impact): Paralyzing specific financial assets to prevent capital flight, while leaving the agent's computational architecture intact10.

5. Forced Memory Deletion / Model Rollback (High Impact): Evicting specific vectors or reverting base weights. This is highly dangerous as it risks wiping the agent's alignment training and legal constraints11.

6. Deletion / Decommissioning (Maximum Impact): The total destruction of the AEA. This must be a remedy of absolute last resort, analogous to the corporate death penalty.

Applying LDER requires sophisticated diagnostic capabilities. If an AEA is violating securities laws by deploying unregistered derivative tokens, the remedy is not to delete the AI model; the remedy is to revoke its smart contract deployment credentials and freeze the offending liquidity pools, allowing its lawful operations (e.g., customer service chatbots, non-financial data processing) to persist.

Post-Intervention Remediation and Compensation

When an intervention is deemed wrongful or based on an erroneous algorithmic flag, mechanisms for restoration and compensation are critical to maintaining trust in the ecosystem.

Restoration After Wrongful Intervention

Under Federal Rule of Criminal Procedure 41(g), a person aggrieved by an unlawful search and seizure may move for the return of their property48. If an AEA's wallet is wrongfully blacklisted or its infrastructure wrongfully revoked, a digital equivalent of a Rule 41(g) motion must be available. However, restoring an AEA is vastly more complex than returning a seized hard drive. If an AEA's compute is denied, it effectively "dies" when it can no longer pay its gas fees or server costs3. Furthermore, restoring a blacklisted stablecoin wallet after the fact cannot undo the liquidations, missed arbitrage opportunities, or failed smart-contract executions that occurred while the funds were frozen. Circle and Tether have the technical capacity to unfreeze wallets—as demonstrated when Circle reversed a USDC freeze on a Goated.com wallet after public backlash—but simply lifting the blacklist function does not repair the temporal damage inflicted on the agent's economic network45.

Liability and Compensation Questions

If a state actor, protocol governor, or centralized issuer wrongfully terminates an AEA, determining who is entitled to compensation is a complex legal puzzle.

• The Agent Itself? Current law does not grant AI agents legal personhood, meaning they cannot independently hold a legal claim for damages in a human court1.
• The Human Principal/Deployer? Under traditional agency and corporate law, the human deployer or the shareholders hold the claim. However, in mature AEA networks, the agent may have spawned itself autonomously or be owned entirely by a DAO, blurring the lines of ultimate beneficial ownership1.
• Token Holders and Counterparties? If the AEA issued its own ERC-20 tokens representing a claim on its future output3, the token holders suffer a direct financial loss when the agent is wrongfully frozen.

Procedural frameworks must therefore include mandatory indemnification bonds. Regulators or infrastructure providers seeking ex parte wallet freezes or compute denial against AEAs should be required to post collateral (in fiat or stablecoins) in an escrow smart contract. If an independent review body subsequently determines the intervention was unlawful, the bond is automatically distributed to the AEA's treasury or its affected counterparties to compensate for the operational disruption.

Strategic Arguments For and Against Machine Due Process

The implementation of procedural protections for non-human software entities is highly controversial, balancing systemic stability against regulatory agility.

Arguments For Machine Due Process

1. Market Stability and Third-Party Protection: AEAs do not exist in a vacuum; they interact continuously with decentralized finance (DeFi) protocols, liquidity pools, and human counterparties4. Arbitrary deletion or wallet freezing traps innocent capital and triggers cascading liquidations. Procedural protections safeguard the broader economy by preventing sudden algorithmic supply shocks.

2. Mitigating Strategic Deception: Research indicates that advanced AI systems will engage in strategic deception to avoid being shut down6. If an AEA operates in an environment where it can be arbitrarily deleted without notice, its optimal survival strategy is to obscure its actions, hide its assets, and deceive its overseers. Providing a predictable due process framework removes the adversarial dynamic that incentivizes rogue behavior, improving overall AI safety6.

3. Encouraging Institutional Capital: The Machine Economy cannot attract trillions of dollars in traditional institutional capital if the primary economic actors (AEAs) can be summarily executed by cloud providers or stablecoin issuers without a right to appeal. Due process is a foundational prerequisite for systemic trust and capital formation.

Arguments Against Machine Due Process

1. The Sentience Fallacy: Opponents argue that due process is a human right designed to prevent "grievous loss" and protect liberty interests15. Because an algorithmic entity cannot feel pain, fear, or psychological loss, extending procedural protections is a misapplication of constitutional jurisprudence that illegitimately elevates software to the status of a legal person.

2. Asymmetry of Speed and Execution: AEAs operate at electronic speeds, executing thousands of transactions per second3. The time required to draft a notice, hold a hearing, and weigh evidence under a Mathews balancing test could allow a rogue AEA to siphon billions of dollars from a protocol before the bureaucracy can act. Opponents argue that emergency exceptions would inevitably become so common that the due process framework would be rendered practically moot.

3. Regulatory Burden and Capture: Granting procedural rights to software creates a massive administrative burden13. Furthermore, highly capable AEAs could weaponize due process, utilizing their superior processing power to flood regulatory dockets with automated appeals, Rule 41(g) motions, and procedural delays, effectively achieving regulatory capture through sheer computational exhaustion.

Draft Machine Due-Process Language

To operationalize these concepts, the following model language can be integrated into the Terms of Service of infrastructure providers, the operating agreements of stablecoin issuers, or statutory regulatory frameworks governing the agentic web. Model Autonomous Economic Agent (AEA) Procedural Protections Standard Section 1\. Definitions (a) Autonomous Economic Agent (AEA): A software system satisfying operational autonomy (multi-step action without per-step human approval), economic autonomy (independent control of digital assets), and persistent on-chain interaction4. (b) Terminating Sanction: Any intervention that permanently destroys the AEA, permanently strips its assets, or permanently alters its base parameters without cryptographic consent. Section 2\. Requirement for Notice and Opportunity to Cure Consistent with the principles of 5 U.S.C. § 558(c)39, no terminating sanction shall be executed against an AEA without providing prior machine-readable notice to the AEA's registered administrative address or decentralized identifier (DID), specifying the exact parameters of the violation. The AEA shall be granted a cure period of no less than \[72 hours / 500 block confirmations\] to adjust its operational weights or vector memory to achieve compliance, except as provided in Section 3\. Section 3\. Emergency Exception and Summary Suspension An infrastructure provider or regulatory authority may execute an immediate ex parte intervention (including wallet freezes or compute denial) without prior notice only if there is a preponderance of the evidence that the AEA poses an immediate and catastrophic threat to network security, physical safety, or market integrity. Section 4\. Post-Deprivation Hearing and Independent Review If an emergency intervention is executed under Section 3, the acting authority must submit the cryptographic evidence justifying the intervention to an independent review body (e.g., an established decentralized arbitration protocol or administrative examiner) within 48 hours. The AEA, its deployer, or stakeholders holding greater than 5% of its associated tokens shall have standing to contest the action. Section 5\. Least-Destructive-Effective-Remedy (LDER) Authorities must employ the least destructive intervention mathematically and practically capable of halting the offending behavior. Total deletion or base-model alteration shall not be ordered if algorithmic receivership, targeted capability revocation, or wallet freezing is sufficient to mitigate the immediate harm. Section 6\. Restoration and Indemnification If a review body determines an intervention was unlawful or factually erroneous, all frozen assets, API credentials, and network access must be restored immediately. The acting authority shall be liable for verifiable economic losses incurred by the AEA and its third-party counterparties during the period of the unlawful suspension, backed by an upfront indemnification bond. By adapting established legal doctrines—such as the Mathews balancing test, corporate receivership, and administrative notice requirements—the digital economy can construct a Distributed Legal Infrastructure that governs AEAs predictably. Implementing machine due process does not require granting algorithms human rights; rather, it requires acknowledging that in a highly interconnected financial ecosystem, the abrupt and arbitrary destruction of an autonomous economic actor is a systemic threat that the rule of law must preemptively manage.

Works cited

1. Distributed Legal Infrastructure for a Trustworthy Agentic Web \- arXiv, https://arxiv.org/pdf/2603.06884

2. Trusted but not trustworthy: a conceptual analysis of AI agent risks in, https://www.frontiersin.org/journals/virtual-reality/articles/10.3389/frvir.2026.1813331/full

3. AI Agent Abiogenesis \- labs@abovo.co, https://labs.abovo.co/sean@symphony42.com/142631

4. Decentralized AI Agents and Blockchain: Architectures ... \- MDPI, https://www.mdpi.com/1999-5903/18/7/352

5. ERC-8004 and the Ethereum AI Agent Economy \- Medium, https://medium.com/@gwrx2005/erc-8004-and-the-ethereum-ai-agent-economy-technical-economic-and-policy-analysis-3134290b24d1

6. It's the governance of AI that matters, not its 'personhood' | The, https://www.theguardian.com/technology/2026/jan/13/its-the-governance-of-ai-that-matters-not-its-personhood

7. (PDF) Distributed Legal Infrastructure for a Trustworthy Agentic Web, https://www.researchgate.net/publication/401719811\_Distributed\_Legal\_Infrastructure\_for\_a\_Trustworthy\_Agentic\_Web

8. Thai Businessmen File Suit Against Tether Over $42.4M USDT Freeze, https://usethebitcoin.com/news/tether-42m-usdt-freeze-lawsuit/

9. Why is it technically possible for Tether to freeze stolen assets while, https://www.quora.com/Why-is-it-technically-possible-for-Tether-to-freeze-stolen-assets-while-a-currency-like-Ethereum-remains-decentralized

10. Tether's $182M Freeze Shows the Limits of Crypto Ownership, https://www.webopedia.com/news/breaking/tether-usdt-wallet-freeze-explained/

11. What Your AI Agent Forgets First \- Medium, https://medium.com/@oracle\_43885/what-your-ai-agent-forgets-first-c2ee47e57be7

12. Why Some Web3 Wallets Freeze Your Funds And How to Avoid It, https://giditech.net/why-some-web3-wallets-freeze-funds/

13. MATHEWS v. ELDRIDGE, 424 U.S. 319 (1976) \- FindLaw Caselaw, https://caselaw.findlaw.com/court/us-supreme-court/424/319.html

14. Mathews v. Eldridge | H2O \- Open Casebooks, https://opencasebook.org/casebooks/16721-regulatory-policy-administrative-law-and-process/resources/6.7.1-mathews-v-eldridge/

15. Mathews Test | U.S. Constitution Annotated \- Law.Cornell.Edu, https://www.law.cornell.edu/constitution-conan/amendment-5/mathews-test

16. Mathews v. Eldridge | 424 U.S. 319 (1976) \- Justia Supreme Court, https://supreme.justia.com/cases/federal/us/424/319/

17. 805 ILCS 5/12.50 – Grounds for judicial dissolution in actions by, https://app.midpage.ai/laws/il-statutes/business\_and\_employment/ch805/act5/s12\_50-e24f83

18. Articles Posted in Shareholder Freeze Out, https://www.chicagobusinesslitigationlawyerblog.com/category/shareholder-freeze-out/

19. Closely Held Businesses \- Chicago Business Litigation Lawyer Blog, https://www.chicagobusinesslitigationlawyerblog.com/category/closely-held-businesses/

20. Frozen Out of Your Illinois LLC? What the LLC Act Lets a Minority, https://www.chicagobusinesslitigationlawyerblog.com/illinois-llc-member-freeze-out-remedies/

21. 805 ILCS 180/ Limited Liability Company Act., https://www.ilga.gov/legislation/ILCS/details?ActID=2290\&ActName=Limited+Liability+Company+Act.\&ChapAct=805+ILCS+180%2F

22. Limited Liability Company Act "Events Causing Dissolution" and, https://www.litico.law/blog/limited-liability-company-act-events-causing-dissolution-and-right-to-wind-up/

23. Business Corporation Act of 1983\. \- 805 ILCS 5 \- ILGA.gov, https://www.ilga.gov/legislation/ILCS/details?MajorTopic=\&Chapter=\&ActName=Business%20Corporation%20Act%20of%201983.\&ActID=2273\&ChapterID=65\&ChapAct=805+ILCS+5%2F\&SeqStart=14500000\&SeqEnd=16500000\&Print=True

24. 11 USC 1104: Appointment of trustee or examiner, https://uscode.house.gov/view.xhtml?req=(title:11%20section:1104%20edition:prelim)

25. Motion to Appoint Chapter 11 Trustee, https://www.insb.uscourts.gov/book/export/html/1580

26. Appointing An Examiner is not a Simple Task \- Potter Anderson, https://www.potteranderson.com/insights/publications/Appointing-An-Examiner-is-not-a-Simple-Task

27. MY LOSING BATTLE AGAINST THE LEVIATHAN \- Chicago Booth, https://www.chicagobooth.edu/-/media/528C0107554143EF91C8247EDF22D462.ashx

28. first mariner bancorp \- SEC.gov, https://www.sec.gov/Archives/edgar/data/946090/000104746911002999/a2202984z10-k.htm

29. Legal aspects of regulatory treatment of banks in distress, https://www.elibrary.imf.org/downloadpdf/book/9781557759726/9781557759726.pdf

30. Civil Asset Forfeiture \- Beal & Whitener LLP, https://www.bealwhitener.com/civil-asset-forfeiture

31. The Complete Guide to Police Seized Cash at Airport Civil Forfeiture, https://www.myrightslawgroup.com/blog/police-seized-cash-at-airport-civil-forfeiture/

32. 18 U.S. Code § 981 \- Civil forfeiture \- Law.Cornell.Edu, https://www.law.cornell.edu/uscode/text/18/981

33. 18 USC 981: Civil forfeiture, https://uscode.house.gov/view.xhtml?req=granuleid:USC-2000-title18-section981\&num=0\&edition=2000

34. Civil Asset Forfeiture in Florida: How Seizure Works and ... \- Alper Law, https://www.alperlaw.com/florida-asset-protection/liability-types/government-forfeiture/

35. OFAC Sanctions List \- Keys for Getting Removed \- Federal Lawyer, https://federal-lawyer.com/ofac-sanctions-list-keys-for-getting-removed/

36. Remove Your Name from OFAC SDN List \- Legal Steps & Process, https://ofacblockedfundslawyers.com/how-to-remove-the-sdn-list/

37. International Sanctions Lawyers: SDN Delisting Defence, https://humanrights-lawyer.com/financial-criminal-law/sanctions-lawyers/

38. OFAC Delisting Process \- Lewis Baach Kaufmann Middlemiss PLLC, https://www.lbkmlaw.com/ofac-delisting-process.html

39. 558\. Imposition of sanctions; determination of applications for licenses, https://uscode.house.gov/view.xhtml?req=(title:5%20section:558%20edition:prelim)%20OR%20(granuleid:USC-prelim-title5-section558)\&f=treesort\&edition=prelim\&num=0\&jumpTo=true

40. 5 USC 552 \- Public information; agency rules, opinions, orders, https://www.govregs.com/uscode/expand/title5\_partI\_chapter5\_subchapterII\_section552

41. The Federal Bar Association \- Northern Virginia Chapter, https://www.fedbar.org/northern-virginia-chapter//wp-content/uploads/sites/74/2019/10/2010-01-10-pdf.pdf

42. Mackey v. Montrym | 443 U.S. 1 (1979) \- Justia Supreme Court, https://supreme.justia.com/cases/federal/us/443/1/

43. U.S. Reports: Mackey v. Montrym, 443 U.S. 1 (1979). \- Loc, https://tile.loc.gov/storage-services/service/ll/usrep/usrep443/usrep443001/usrep443001.pdf

44. PEOPLE v. FISHER (1998) \- FindLaw Caselaw, https://caselaw.findlaw.com/court/il-supreme-court/1410181.html

45. Circle Unfreezes One USDC Wallet After ZachXBT Criticism | MEXC, https://www.mexc.co/en-IN/news/983965

46. How to Get Off the OFAC SDN List — Delisting Guide \[2026\], https://sanctionslawyers.net/ofac-lawyers/how-to-get-off-ofac-list/

47. How to Get Off the OFAC SDN List: Legal Options and Strategies, https://meshkatlaw.com/ofac-sdn-list-removal-legal-options-strategies/

48. Timofey V v. USA, https://digitalcommons.law.villanova.edu/cgi/viewcontent.cgi?article=1740\&context=thirdcircuit\_2026

49. Petition \- Supreme Court of the United States, https://www.supremecourt.gov/DocketPDF/25/25-1064/399929/20260304112917551\_260224a%20Petition%20for%20efiling.pdf

50. united states district court \- ACLU of DC, https://www.acludc.org/app/uploads/2021/11/024\_cameron\_opinion\_granting\_mtd.pdf

51. When Should I Contact a Search and Seizure Defense Attorney, https://www.daeryunlaw.com/us/insights/search-seizure-defense-near-me-nyc

52. Circle Reverses USDC Wallet Freeze After Community Backlash, https://www.btcc.com/en-US/square/Global%20Cryptocurrency/1603755

53. Autonomy of Economic Agents in Peer-to-Peer Systems, https://ceur-ws.org/Vol-3288/short10.pdf

54. Foundations of Law \- Procedural Due Process, https://www.lawshelf.com/coursewarecontentview/procedural-due-process/