Regulatory Triggers
This section specifies the minimum conditions under which action is no longer discretionary. When these conditions are met, continued inaction constitutes a failure of duty of care. The purpose is activation, understood as the withdrawal of discretionary innocence under known operating conditions.
I. Trigger Condition A: Systemic Risk is Established
Condition:
Harm recurs across time, platforms, and jurisdictions despite removal of individual actors or content.
Evidence Standard:
Reappearance through substitution (new accounts, mirrors, forks).
Persistence under enforcement pressure.
Predictable escalation pathways documented across cases.
Implication:
The phenomenon is structural, not episodic. It qualifies as systemic risk.
Action Obligation:
Regulatory focus must shift from instance-level moderation to architecture-level intervention.
II. Trigger Condition B: Selection Effects Are Operating
Condition:
System optimization criteria (e.g., engagement, retention, growth) differentially amplify harmful material because it outperforms alternatives on platform metrics.
Evidence Standard:
Harm correlates with amplification signals (watch time, repost velocity, reaction density).
Removal does not reduce exposure pathways. Recommendation systems continue to surface adjacent material.
Changes in optimization parameters plausibly change prevalence.
Implication:
Harm is selected, not incidental.
Action Obligation:
Platforms must alter optimization targets, thresholds, or routing logic where foreseeable amplification occurs.
III. Trigger Condition C: The Useless Machine State is Reached
Condition:
Governance mechanisms (reporting tools, moderation pipelines, trust-and-safety processes) process increasing inputs without reducing harm prevalence or circulation velocity.
Evidence Standard:High report volumes with negligible impact on recurrence.
Procedural completion metrics improve while harm metrics remain unchanged.
Burden of response falls primarily on reporters and moderators.
Implication:
Governance has entered a stable failure equilibrium: a useless machine.
Action Obligation:
Procedural reform alone is insufficient. Upstream constraints must be imposed on amplification and monetization.
IV. Trigger Condition D: Duty of Care Is Foreseeably Unmet
Condition:
Risk pathways are documented, mitigation is technically feasible, and no proportionate design change is implemented.
Evidence Standard:
Known escalation gradients and monetization routes remain intact.
Interventions remain downstream of amplification.
Platforms acknowledge harm but maintain unchanged architectures.
Implication:
Responsibility attaches to system operators, independent of intent.
Action Obligation:
Failure to act constitutes breach of duty of care.
V. Consolidated Regulatory Trigger
When Conditions A–D are simultaneously satisfied, the following conclusions are compelled:
1. The harm qualifies as systemic risk.
2. The system is selecting for harm through its optimization logic.
3. Existing governance operates as a useless machine.
4. Duty of care is breached by continued architectural inaction.
At this point, discretion ends.
VI. Required Intervention Classes (Non-Exhaustive)
Visibility reduction alone is insufficient. Without prescribing implementation detail, effective action must target structure, not sentiment.
Amplification Constraints
Pre-distribution throttling where risk is foreseeable.
Monetization Disruption
Removal of economic incentives that reward circulation.
Friction Reintroduction
Cost imposition on repeated harmful propagation.
Metric Realignment
Success criteria tied to harm reduction, not throughput.
VII. Closure
These conditions do not require moral consensus, ideological alignment, or expanded evidentiary discovery. They require recognition of what the system already demonstrates.
Where harm persists predictably, scales reliably, and resists procedural remedy, action is no longer optional. The obligation follows from structure.
On the framework established here, the conditions are met in principle and warrant supervisory inquiry, audit escalation, and routing-layer scrutiny. The framework is intended as a transfer structure, not a claim of perfect equivalence across harm domains.
Attribution Closure Test (Regime Shift Sufficiency Conditions)
Purpose
To distinguish symbolic attribution (stabilizing) from enforceable attribution (destabilizing).
If any condition fails, the system remains metastable.
Condition 1: Causal Necessity Is Legible
Test
Can persistence at scale be shown to require a specific system function?
Pass Criteria
The harm does not aggregate, stabilize, or reproduce without recommender systems, ranking heuristics, cross-surface relay (e.g., feed ? groups ? semi-closed), or equivalent routing mechanisms. Removal or neutralization of the routing function produces a measurable collapse in scale, persistence, or replication velocity.
Fail Indicators
Harm persists via “user choice,” manual search, or isolated uploads without routing assistance.
Platform claims the system merely “hosts” or “reflects” demand.
Interpretation
Routing is not transport: it is the enabling condition.
If routing necessity cannot be demonstrated, attribution cannot attach.
Condition 2: Structural Enforceability Attaches
Test
Does enforceable cost land on the routing layer itself?
Pass criteria
Liability, injunction, or binding duty targets: ranking parameters, recommendation objectives, amplification thresholds, or discovery mechanics.
Enforcement cannot be satisfied by content removal, account bans, policy statements, staffing increases, or reporting obligations alone.
Fail Indicators
Enforcement focuses on items, users, or moderation throughput.
Compliance is satisfied through process, not architectural change.
Interpretation
If routing economics remain intact, enforceability has not attached.
Condition 3: Attribution Ambiguity Collapses Across the Chain
Test
Is the actor with capacity to prevent persistence non-evadable?
Pass criteria
Responsibility cannot be displaced onto users, moderators, NGOs, journalists, jurisdictions, or downstream institutions.
Capacity + foreseeability + persistence are sufficient for responsibility.
Cross-border design does not defeat enforceability.
Fail Indicators
Responsibility fragments across interfaces or jurisdictions. Platforms invoke privacy, encryption, or neutrality to deny capacity.
Interpretation
Where capacity exists and persistence is foreseeable, ambiguity is a design artifact.
Condition 4: Anti-Replication Is Enforced
Test
Can replication outpace intervention after enforcement?
Pass criteria
Altered routing parameters reduce re-upload viability, mirroring advantages, format migration success. Substitution no longer restores prior scale.
Fail Indicators
New accounts, formats, or surfaces rapidly restore circulation.
Enforcement lags remain structurally inevitable.
Interpretation
If replication outpaces enforcement, attribution has not closed.
Decision Rule
All four conditions satisfied ? Phase transition possible.
Any condition unsatisfied ? Recognition remains detection. Persistence continues.
Non-Conversion Set (Explicit)
The following do not constitute attribution closure on their own:
Visibility or exposure
Condemnation or outrage
Post-hoc takedowns
Account bans
Policy expansion
Transparency reports
Oversight boards
NGO or trusted-flagger partnerships
Increased moderation headcount
AI detection improvements without routing change
Each is compatible with persistence.
This work is licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0).
License: https://creativecommons.org/licenses/by/4.0/ © 2026 Elijah Morgan
DOI: https://doi.org/10.5281/zenodo.19599231
Published 2026-04-15 Version 1.1