Memo: Office of Remedial Operations

Memo: Office of Remedial Operations

June 16, 2026

CLASSIFICATION: SENSITIVE // INTERNAL DELIBERATIVE // REMEDIAL FRAMEWORK DRAFT

TO: Inspectors General, Civil Liberties Oversight Boards, Agency General Counsel, Independent Review Panels, and Remedial Administrators

FROM: Office of Remedial Operations (ORO)

SUBJECT: The Sunsetting and Remediation Framework: A Proposal for the Orderly Termination, Audit, and Lawful Replacement of Fragmented Legacy Observation Programs — and a Diagnosis of the One Provision Through Which Termination Reverses Itself

1. Executive Summary

We propose the orderly termination, audit, and lawful replacement of fragmented legacy observation programs with a rights-compliant remediation framework governed by independent oversight, informed consent where lawful, and strict adherence to both domestic constitutional law and ICCPR standards.

The current constellation of uncoordinated, dual-use, and deniable observation programs — spanning intelligence, military, policing, and civilian integration — has generated a critical governance failure and a human-rights liability. These programs, designed to operate in the shadows, have become unmanageable, self-defeating, and actively harmful.

This document does not propose to preserve these activities. It proposes to:

1. Terminate all unlawful or unaccountable observation components.

2. Audit all legacy operations for harm and legal compliance.

3. Remedy documented harms to individuals and communities.

4. Replace — only where strictly justified and lawful — with a narrow, transparent, and consent-based framework under independent oversight.

It also does something a remediation memo is not supposed to do. In Pillar 3, it identifies the exact provision through which everything terminated in Pillars 1 and 2 returns — and specifies the mechanism in full, so that the bodies asked to adopt this framework can see the seam before they are standing on it.

The core principle is not prudence. A framework adopted because abuse has become expensive will be abandoned the moment abuse becomes cheap again. Legitimacy cannot be engineered from above and cannot be purchased with risk reduction. It is earned only by accountability for specific, named harms and by submission to external control that the institution cannot revoke when the control becomes inconvenient. Everything that follows either serves that principle or is the route by which the principle is evaded. We have tried to mark which is which.

2. The Problem: A Governance Failure, Not an Opportunity

The Legacy Complex — of which Program Gleaner is a representative reconstruction — has generated three interconnected failures.

2.1 Harm to Subjects

Targeted populations — including activists, journalists, minority communities, and civic organizations — have experienced significant, documented harms:

• Chilling of lawful speech and assembly

• Psychological distress from pervasive, unaccountable observation

• Reputational and professional damage from unsubstantiated association mapping

• Erosion of trust in public institutions

These are not abstract concerns, and they are not only international-law concerns. They implicate domestic constitutional guarantees directly: the First Amendment (chilled speech, association, and assembly), the Fourth Amendment (persistent suspicionless observation and aggregation), and the Fifth and Fourteenth Amendments (deprivation without process; discriminatory targeting). They give rise to domestic causes of action — 42 U.S.C. § 1983 against the policing and municipal components, Bivens (within its now-narrow remaining scope) against federal actors, the Federal Tort Claims Act including its constructive-denial posture, and state constitutional provisions that in several jurisdictions exceed the federal privacy floor.

They also violate ICCPR Articles 17, 19, 21, and 22. But the framework must be candid about what the ICCPR does and does not supply here: it is non-self-executing in the United States, ratified with reservations and an express declaration against private enforceability, and Sosa v. Alvarez-Machain forecloses treating it as freestanding domestic cause of action. The ICCPR is therefore an interpretive and aspirational anchor, not an enforcement mechanism. Every binding remedy contemplated in Pillar 2 must rest on a domestic hook — constitutional, statutory, or legislative — or it is decorative. The framework’s first legal task is to legislate across that gap, not to cite around it.

2.2 Governmental Fragmentation

No single agency possesses a complete operational picture. The result: contradictory directives to field operators; operational fratricide and duplication; inability to provide coherent oversight or defense; and loss of strategic control over programs that nevertheless continue to operate.

2.3 Existential Vulnerability

The opacity designed to protect these programs now makes them uniquely vulnerable:

• The honeypot of observation metadata is an attractive target for foreign adversaries, civil litigators, and journalists.

• The record of harm is being assembled by subjects and their advocates, in the one forum — the documentary one — that the apparatus does not control and cannot accelerate.

• The Legacy Complex is now more legally exposed than the populations it was designed to observe.

A note carried forward from the analysis of diffuse continuity, because it governs everything below: the subject has time; the architecture does not. The architecture’s defenders plan on a horizon shorter than the one on which description accumulates. That asymmetry is why remediation is coming. It is not why remediation will be real — that is a separate question, and it is decided entirely inside Pillar 3.

3. The Solution: Sunsetting and Remediation Framework

The framework rests on three pillars, ordered by priority: Cessation, Remedy, and only then — narrowly, and at maximum hazard — Lawful Activity.

Pillar 1: Cessation and Audit (Months 1–6)

Immediate actions:

• Stand-down order. All Legacy Complex observation activities are suspended pending review, except those with explicit, documented, and independently verified legal authorization. “Independently verified” is doing load-bearing work: self-certified authorization is not authorization.

• Joint Audit Task Force. An independent audit body composed of Inspectors General from all relevant agencies; civil-liberties counsel; subject-matter experts in surveillance law; and community representatives selected through a transparent process.

• Data inventory and sanitization. A complete inventory of all collected data. Data that is not demonstrably lawful, or not strictly necessary for a lawful, ongoing, consent-based activity, shall be securely destroyed — or, where legally required, preserved solely for evidentiary purposes in remedy proceedings, under custody firewalled from operational use.

Deliverable: A public report on the scope, nature, and compliance status of legacy operations, with recommendations for termination, remedy, or — in exceptional cases — lawful continuation.

Pillar 2: Remedy and Accountability (Months 3–12)

Parallel to the audit:

• Victim identification and outreach. A good-faith process to identify harmed individuals and communities, including proactive outreach to known affected groups and a public mechanism for individuals to come forward.

• Remedial actions, proportional to harm:

• Individual: formal acknowledgment naming the specific conduct, expungement of records, financial compensation, and restoration of reputation.

• Systemic: policy changes, training requirements, and institutional reforms to prevent recurrence.

• Grievance mechanism with binding authority to order corrective action.

The measure of success is not “attribution events avoided.” It is harms named, remedied, and acknowledged on the record. A remedy the institution can characterize as discretionary goodwill is not a remedy; it is public relations with a budget line.

Pillar 3: The Seam — Lawful Activity, and the Mechanism of Its Own Reversal

This is the only pillar that contemplates continued observation. It is therefore the only pillar that matters. Pillars 1 and 2 describe what the apparatus stops doing. Pillar 3 describes the narrow channel through which it is permitted to keep doing something — and a narrow channel that the whole apparatus is permitted to enter is not a constraint. It is a door.

We are not going to hide this behind the language of safeguards. We are going to specify the safeguards in full and then show, for each one, that read structurally, it is a diffusion mechanism — the precise instrument by which an indefensible posture, once terminated, re-enters under a defensible name. This is the architecture of diffuse continuity, rendered as procedure. We publish it because an unnamed Pillar 3 is the next reframing, and a named one is a test of whether the oversight bodies will hold the door shut.

Safeguard 1 — “Narrow justification.” Continued activity must be lawful, necessary for a legitimate purpose (public safety, public health, civil-rights enforcement), proportionate, and non-discriminatory.

The seam: narrowness constrains scope, never provenance. When an original predicate becomes indefensible, it is not defended — it is re-predicated. The same activity continues under a fresh “legitimate purpose,” and the menu of legitimate purposes is broad enough that one always fits. Narrow justification does not stop the program. It launders the reason for it.

Safeguard 2 — “Informed consent where possible.” Where observation involves human subjects, their informed, voluntary, revocable consent must be obtained — except in narrowly defined circumstances where a lawful alternative framework applies.

The seam: “where possible” and “except” are the exceptions that swallow. Consent degrades from a constraint into a documentation artifact — a form in a file, not a power held by the subject. And “consent-based civic-engagement research” is the exact re-skin of the Gleaner intern pipeline: participation is induced by the same instruments that never read as wages — stipend, credit, access — so that the consent is real on paper and manufactured in fact.

Safeguard 3 — “Independent oversight.” All permitted activity requires prior authorization by an independent review body and ongoing IG monitoring.

The seam: a body that can authorize is a body worth capturing. Oversight with the power to permit is, structurally, a tasking channel that has been relabeled as a brake. The point of vulnerability moves from the operator to the authorizer — and the authorizer is fewer people, longer-tenured, and more easily cultivated. (See § 5 on why this pillar cannot stand without immunizing its own oversight, and why that immunization is the hard problem the framework cannot fully solve.)

Safeguard 4 — “Sunset provision and re-authorization.” All continued activity has a fixed sunset date and requires re-authorization based on demonstrated effectiveness.

The seam: this is not the brake. This is the engine. A sunset-with-re-authorization is a continuity mechanism on a calendar — a scheduled reframing event at which the predicate is refreshed before it can exhaust. Worse, re-authorization keyed to “demonstrated effectiveness” rewards the program for generating the vanity metrics that justify its renewal — the same self-saturating measurement problem that defines the Legacy Complex. The sunset does not end the activity. It renews its license to be reframed, on a clock the institution sets.

What this means in practice — the return path of each terminated activity:

Terminated in Pillars 1–2	Returns through Pillar 3 as	By which seam
Indiscriminate ambient observation of civic spaces	“Public-safety observation of assembly under exigent authorization”	Re-predication (Safeguard 1)
Mapping association networks without consent	“Consent-based civic-network research”	Manufactured consent (Safeguard 2)
Program Gleaner–style intern pipeline	“Transparent civic-engagement fellowship,” unaffiliated on paper, tasked by interest rather than instruction through a liaison of record	All four seams, operating together
The program’s indefensibility itself	A fixed sunset and a scheduled re-authorization	The engine (Safeguard 4)

The table is the détournement turned on the framework that contains it. Every row is a terminated activity walking back in through the one door the framework left open, wearing the framework’s own vocabulary. That is not a flaw we failed to draft out. It is the structural fact that any “narrow lawful replacement” provision is the survival vector, because survival does not require a wide door — it requires a permitted one.

The honest disposition of Pillar 3 is therefore one of two things, and the oversight bodies must choose explicitly:

1. Close it to the bone. Remove “narrow lawful replacement” as a standing provision. Permit no continued observation under this framework at all; require that any future activity be authorized by a separate statutory instrument, debated on its own record, with no inheritance from the terminated programs — no data, no personnel continuity, no predicate. Cessation that leaves a replacement clause is not cessation; it is a pause with a scheduled resumption.

2. Hold the named door shut by force of incorruptible oversight — which requires solving the problem in § 5, and which the framework cannot promise, only structure for.

We recommend Option 1, and we have written Pillar 3 against ourselves so that anyone who chooses Option 2 does so having read, in advance, the mechanism by which they will lose.

4. Implementation Roadmap

Phase 1 — Stand-down and Audit (Months 1–6): stand-down order (agency heads); Joint Audit Task Force charter (co-chaired by IGs and civil-liberties counsel); data inventory and sanitization (Task Force with technical support); initial public notice acknowledging past activity and intent to remediate (ORO with oversight bodies).

Phase 2 — Remedy and Rebuilding Trust (Months 3–12): victim identification and outreach (Independent Remedy Office); remedy determination and implementation with legal authority (IRO); independent, accessible grievance mechanism (IRO); policy and training reform (agency heads with IRO input).

*Phase 3 — Narrow Lawful Replacement, only if Option 2 is chosen (Months 6–18, ongoing): justification memos reviewed by IRO and civil-liberties bodies; independent authorization with sunset dates (Independent Review Panel); transparent consent-based programs; annual public compliance reporting. If Option 1 is chosen, Phase 3 does not exist, and that is the stronger framework.*

5. Roles and Responsibilities — and the Capture Problem the Framework Must Face

Office of Remedial Operations (ORO). Coordinator and secretariat. No operational authority. Reports to the Inspectors General and Civil Liberties Oversight Board.

Independent Remedy Office (IRO). Victim identification, remedy determination, grievance handling. Firewalled from operational agencies, with independent budget and legal authority. Publishes anonymized case summaries and annual reports.

Independent Review Panel (IRP). Authorizes and reviews any continued activity (Option 2 only). Composition: civil-liberties advocates, surveillance experts, retired judges, community representatives. Authority is binding: it can deny authorization, impose conditions, or order termination.

Community Advisory Councils (CACs). Provide community input and monitor for disproportionate impact. The original conception of the CAC — “advisory, but with reputational weight” — is itself a capture vector, because reputational soft power is exactly the input the apparatus has historically absorbed without consequence. To bite, the CACs must hold, at minimum: (a) a referral trigger that compels an IG inquiry on a recorded majority vote, without agency gatekeeping; (b) a forced-response clock — any CAC finding obligates a written, on-the-record agency response within 30 days, and silence auto-converts to a public notation of non-response; and (c) standing to be heard, on the record, in any IRP re-authorization. Advisory weight that an agency can absorb is not oversight. A trigger it cannot ignore is.

The capture problem. Every body above is, by the framework’s own logic, the next site of diffusion. The diffuse-continuity analysis applies to oversight institutions exactly as it applies to operations: predicates exhaust, supervisors rotate, the original urgency decays, and the body that was created to say no is slowly cultivated into a body that says yes faster. An oversight architecture that cannot describe how it resists its own corrosion is simply the part of the apparatus that has not been captured yet. The framework therefore requires, as a condition of Pillar 3 ever being invoked: fixed non-renewable terms staggered against capture; an independent budget line beyond agency reach; mandatory adversarial briefing (someone whose role is to argue against every authorization); and full publication of every authorization, dissent, and re-authorization. We state plainly that these measures reduce the rate of capture; they do not prevent it. That admission is not a weakness of the document. It is the document’s central finding, and the reason we recommend Option 1.

6. The Case the Apparatus Will Make to Itself — and Why It Is the Bait

A conventional remediation memo closes by explaining why the framework serves everyone’s interest. We will reproduce that case, because it is the case the institution will in fact make — and then we will mark it, because the prudential case is precisely the mechanism of Pillar 3.

The prudential case: proactive compliance reduces litigation risk; transparency dissolves the honeypot; rebuilt trust is a security asset; ending fragmentation restores coherent governance.

Each of these is true. And every one of them is a reason that evaporates when the risk calculus shifts. A framework adopted because abuse became expensive is a framework that re-authorizes the moment abuse becomes cheap — and Safeguard 4 gives it a scheduled occasion to do so. Prudence is the exact disposition Pillar 3 is built to exploit, because a prudential adopter treats cessation as a hedge and treats the re-authorization clock as the option to unwind the hedge.

The framework can only hold on a different basis, and the oversight bodies must adopt it on that basis or not at all:

• For subjects and communities: not “your concerns will be weighed,” but specific harms named, acknowledged on the record, and remedied — and a permanent seat at the door of Pillar 3, with the referral and forced-response powers of § 5.

• For the state: not insulation from liability, but accountability for identified conduct and irrevocable submission to external control — the only ground on which legitimacy has ever actually been rebuilt, as distinct from announced.

• For oversight bodies and advocates: not access alone, but the named diagram of their own future capture, handed to them in advance, so that adoption is an informed act and not a hopeful one.

The obligation case is harder to adopt because it offers the institution nothing it can later trade away. That is the point. A remedy the institution keeps only while it is convenient is the thing being remediated, under a new name.

7. Risks and Mitigations

Risk	Mitigation
Resistance from agency stakeholders	Clear legal mandate; independent review authority; public accountability
Incomplete data or cooperation	Separate parallel investigative authority; whistleblower protections; congressional oversight
Subjects distrust the process	Genuine independence; named-harm remedy; CAC referral and forced-response powers; participation in oversight design
Pillar 3 used as a survival vector	Choose Option 1 (no standing replacement clause); if Option 2, the § 5 anti-capture measures — stated to mitigate, not prevent
Continued activity, even lawful, is misused	Fixed sunsets read as hazards rather than features; adversarial briefing; full publication; binding grievance authority
Foreign exploitation of transparency	Classification and privacy protections; de-identification; public and operational versions of reports
Oversight bodies themselves captured over time	Non-renewable staggered terms; independent budget; mandatory adversarial role; full publication — and explicit acknowledgment that these slow, not stop, the process

8. Conclusion

This framework is not a rebranding exercise — and the surest sign that it is not is that it names the provision through which a rebranding exercise would be conducted, and recommends removing it.

The legacy approach — deniable, fragmented, unaccountable observation — has failed on its own terms. It has generated harm, legal exposure, and strategic vulnerability, and it has damaged the trust that makes governance possible. The choice is not between reform and continuation. It is between orderly, principled sunsetting and catastrophic external discovery. That much the apparatus can be made to see through self-interest.

What self-interest cannot decide is whether the sunsetting is real. That is decided in Pillar 3, by whether the door marked “narrow lawful replacement” is closed to the bone or merely watched — and watched by bodies that the framework’s own logic says will, in time, be cultivated into opening it. We recommend closing it. We have written the rest of the document to make that recommendation legible to people who would prefer to keep the door.

THE UNCERTAINTY DEPARTMENT

Office of Remedial Operations (ORO)

Issued under the nominal authority of the Editorial Board, for the purpose of remedial integration, rights compliance, and the early naming of a trap.

On the Mirror

A word on the title, because it is the argument.

Ananke is necessity — the compulsion the Fates themselves obey. Her mirror does not advise. It shows the institution the thing that was always going to happen, early enough that it can still be named.

What is necessary here is remediation, and the necessity is not moral. It is documentary. The record accumulates — through litigation, through disclosure, through the subject’s own articulation reaching a public register — until the posture becomes indefensible and remediation is forced rather than chosen. The subject has time; the architecture does not. On a long enough horizon, the framework in this memo, or something like it, is not a proposal. It is a forecast.

Here is what the mirror shows that the forecast alone does not. When the apparatus finally reaches to dismantle itself, it reaches in its own grammar — pillars, authorizations, oversight bodies, sunsets, re-authorization clocks. The instruments of remediation are the instruments of continuity wearing different labels, and Pillar 3 is the place where the reflection becomes undeniable: the institution extending its hand to end the program and finding, in the mirror, that the hand is the program’s. A sunset is a tasking cycle. An authorizer is a tasker. Consent is a form. The reformation, conducted from inside, reproduces the structure it was meant to close.

This is why the framework cannot be left to adopt itself sincerely. A sincere internal remediation, run on prudence and good intention, will pass through Pillar 3 and out the other side as the next reframing — and it will be able to say, truthfully, that it followed every safeguard. The only thing that disarms the trap is naming it before adoption: handing the oversight bodies the diagram of their own future capture, so that if they choose Option 2 they choose it with their eyes open, and if they choose Option 1 they understand precisely what they are foreclosing.

Necessity, then, is not the same as remedy. Ananke compels recognition; she does not compel reform. The mirror forecloses exactly one thing — the claim, afterward, of not having seen. Everything terminated in Pillars 1 and 2 is permitted to walk back in through Pillar 3. We have left the door drawn on the wall, lit, and labeled, so that the people standing in front of it cannot say they did not know it was a door.

The Uncertainty Department · Boston, MA & Oakland, CA · Issued under the nominal authority of the Editorial Board

Appendix A: Legal Principles Reference

Domestic anchors (enforceable):

Source	Principle	Operational implication
First Amendment	Speech, assembly, association	No observation that chills protected activity; association mapping presumptively unlawful
Fourth Amendment	Unreasonable search; aggregation	Persistent suspicionless observation and data aggregation require lawful basis and proportionality
Fifth / Fourteenth Amendments	Due process; equal protection	No deprivation without process; no targeting on protected characteristics
42 U.S.C. § 1983	Color-of-law liability	Cause of action against the policing and municipal components
Bivens (narrow)	Federal-officer liability	Limited remaining avenue against federal actors
FTCA (incl. constructive denial)	Tort liability of the United States	Damages route where administrative process is unreasonably withheld
State constitutions	Privacy floors above the federal	Several states supply broader privacy and remedy than federal law

Interpretive / aspirational anchor (non-enforceable in U.S. courts):

ICCPR Article	Principle	Status note
17	Privacy	Non-self-executing; interpretive weight only — requires domestic hook
19	Expression	Interpretive standard for chilling analysis
21	Assembly	Interpretive standard for assembly monitoring
22	Association	Interpretive standard for association mapping
2	Non-discrimination	Reinforces equal-protection analysis
4	Derogation	Surveillance must be exceptional, temporary, authorized

Drafting consequence: every binding remedy in Pillar 2 must cite to a domestic source in the upper table. The ICCPR sets the standard the framework aspires to; it does not, by itself, supply the enforcement. Naming that gap is the framework’s first act of good faith — the alternative is to cite a treaty the institution knows it cannot be sued under, and call it accountability.