Who can enforce an RIR public-interest promise?

Summary

Public-interest language can guide directors, constrain corporate purpose, inform interpretation and support regulatory scrutiny, but those functions are different from a private right to challenge a registry decision.
Enforceability requires four linked elements: an eligible claimant, a sufficiently determinate standard, a forum independent enough to test the decision and authority to grant a remedy that changes the legal or operational position.
RIR members often possess concrete voting, reporting, recall or arbitration rights. Downstream customers and other non-members may bear equal or greater harm while lacking access to those corporate and contractual mechanisms.
Recognition criteria and technical dependence make public-interest promises institutionally significant, yet neither automatically creates a global regulator, a final appellate body or a right to compensation.
A credible regime would publish complaint denominators and outcomes, separate review from management, permit urgent preservation, state the standard of review and connect findings to correction, reconsideration, governance action or regulatory referral.

The clause sounds strongest where the remedy is weakest

Public-interest language is one of the most attractive features of Internet governance documents. It tells users that a private corporation entrusted with globally important coordination will not behave like an ordinary vendor pursuing only its own income. It signals that directors should consider technical stability, fair access, competition, community participation and consequences beyond the membership roll. It offers a moral answer to the question of why a singular registry should be trusted.

The language becomes hardest to test at the moment it matters. A resource holder is denied a transfer. A registration changes after contested evidence. A security service is suspended. A fee structure falls heavily on one class. A customer suffers because its provider's upstream relationship fails. Each affected person can say that the outcome harms the public interest. The institution can reply that conserving resources, enforcing policy, protecting security or preserving financial stability also serves the public interest.

If the clause contains no method for choosing between those claims, it does not decide the dispute. It supplies a vocabulary in which both sides can describe themselves as guardians of the same abstraction. The board retains the first and often final interpretation of the promise it is accused of breaking.

That does not make the clause meaningless. Corporate purpose can limit how assets are used. Directors may need to act consistently with the organisation's entities and duties under governing law. Regulators may rely on declared purposes when evaluating status or conduct. Courts may use the clause to interpret an ambiguous power. Members may invoke it politically when electing or removing governors.

But these effects must not be confused with direct enforcement by an affected person. Accountability requires a chain: someone eligible to complain, a standard capable of application to evidence, a forum able to decide independently and a remedy capable of changing the result. Remove any link and the promise mainly operates through reasons, elections, supervision and reputation.

The question is therefore not whether RIRs use noble words. It is who can make those words costly to disregard. A promise that can never support correction may improve speeches and annual reports while leaving the distribution of power unchanged.

Plaintiff, standard, forum and remedy

The first element is the claimant. Corporate law may give members, directors, the corporation itself, a public official or another defined person standing to enforce particular duties. A service agreement may give a holder access to arbitration. A statute may protect consumers, competitors, data subjects or another class. The general public is not automatically one legal person capable of suing whenever an institution invokes public purpose.

The second element is the standard. “Act in the public interest” is too open unless the reviewer knows what factors count, which are mandatory, what evidence is required and how conflicts are balanced. A standard can remain flexible, but it must distinguish permissible judgment from abuse. Relevant tests might include legality, consistency with corporate purpose, procedural fairness, evidence, equal treatment, proportionality, technical necessity and consideration of foreseeable third-party effects.

The third element is the forum. A board reviewing its own decision may correct obvious error, but it does not supply the same assurance as an independent panel, arbitrator, court or regulator. Independence is not a ceremonial label. It concerns appointment, tenure, conflicts, access to records, power to hear affected parties, publication of reasons and protection from retaliation or budget pressure.

The fourth element is remedy. A complaints office can listen without being able to preserve service, correct a record, require reconsideration, award relief, remove an officeholder or refer a breach. A finding that lessons were learned may have reputational value, but it does not restore a route-security entity or reverse an unlawful decision. The remedy must match the harm and arrive while it can still matter.

These elements are cumulative. A strong standard is useless if no one may invoke it. Broad standing is cosmetic if the forum applies no reviewable test. An independent decision is advisory if management can ignore it. A powerful remedy is dangerous if the standard is vague and evidence inaccessible.

RIR documents often contain pieces of this chain in different places. Bylaws give members votes. Service agreements define bilateral duties. policy procedures create appeals. corporate law supplies court supervision. recognition criteria establish institutional expectations. The accountability error is to point to the collection and assume it forms one seamless appeals ladder. Each mechanism has a different claimant, subject and remedy.

ARIN's public-interest promise is a purpose, not a cause of action

The ARIN Articles of Incorporation contain the clearest phrase. The published consolidated text records the corporation's formation in 1997 and states educational, charitable and scientific purposes. Among a longer list, it describes promoting the expansion and growth of Internet infrastructure for the general public and members by means consistent with the public interest.

That clause performs real charter work. It locates the corporation within stated nonprofit purposes. It tells directors that the corporation's funds and powers are not available for any project they prefer. It can inform interpretation of more specific authority. It can matter to officials responsible for company or tax law. It also gives members and the public a benchmark for criticism.

What it does not do on its face is identify a private plaintiff. The general public is named as a beneficiary of activity, but a beneficiary description is not necessarily a grant of litigation rights. The clause does not define a test for deciding whether a transfer policy, security suspension, fee or registry correction is consistent with public interest. It does not name a tribunal. It does not state whether the remedy is an injunction, a new board decision, damages, removal or regulatory action.

The problem is not unique to ARIN, and it should not be resolved by pretending that every purpose clause is legally inert. The effect depends on Virginia corporate law, federal tax status, the challenged act, the claimant and the relief requested. A member contesting corporate procedure presents a different case from a downstream customer seeking compensation for lost service. A public official examining misuse of nonprofit assets stands differently from a competitor objecting to registry conduct.

The clause should therefore be described accurately: it is a corporate purpose with potential interpretive, fiduciary, supervisory and political consequences. Whether it is enforceable in a particular dispute requires a separate legal route. That answer is less dramatic than calling ARIN legally bound to the entire public, but it gives affected parties a usable map.

The institutional danger appears when the broad clause is used asymmetrically. It is cited to justify expansive discretion because almost any registry action can be narrated as protecting the Internet, then treated as too abstract for anyone to enforce when the action is challenged. A purpose that enlarges authority but never supports review is not balanced governance.

The board is both steward and first interpreter

The ARIN Bylaws state that the Board controls the corporation's power, authority, property and affairs subject to the Articles and Virginia law. They also commit ARIN to open, transparent multistakeholder registry policy development and define the rights of Trustee, General and Service Members.

In ordinary operation, the Board must interpret the public purpose. Budgets, reserves, service investments, litigation decisions, policy implementation and risk controls all require judgments about institutional mission. No external tribunal can pre-approve every choice. Board stewardship is therefore necessary.

The accountability issue is what happens when the Board's own interpretation is disputed. If it says an action serves the public interest, what record must it produce? Must it identify affected non-members? Must it consider less harmful alternatives? Must it explain why member benefit does not displace wider responsibility? Can a reviewer test the evidence or only check whether the Board used the correct words?

Broad discretion can be legitimate where technical trade-offs are complex, but deference should attach to demonstrated reasoning rather than institutional status. A reasoned decision would identify the authority used, the factual record, the interests affected, the chosen standard, the alternatives considered, the expected harm and the review route. A bare invocation of stability or community interest should not close the inquiry.

Conflicts sharpen the problem. Board members may be elected through a membership whose interests overlap with, but are not identical to, customers and the wider public. This does not make elected trustees illegitimate. It means conflict rules and reasons must show how the wider purpose was considered when member and non-member interests diverge.

The Board's dual role also affects emergency decisions. Security or continuity may require fast action before full participation. A credible regime would allow temporary measures while requiring prompt independent review, disclosure of the evidentiary basis to the extent possible and a sunset unless the decision is confirmed. Emergency power should not turn public interest into an unreviewable incantation.

The strongest boards welcome this discipline because it separates stewardship from self-certification. They remain responsible for the first decision but do not possess the last word merely because the charter entrusted them with corporate affairs.

Member power is concrete but does not equal public enforcement

ARIN's bylaws give General Members in good standing voting rights and a mechanism to initiate trustee recall. A petition signed by at least 10 percent of that member class can trigger the stated process, and removal depends on a member vote subject to quorum and election rules. These provisions identify a claimant class, threshold, procedure and governance remedy.

That is genuine accountability. A trustee knows that members can remove them. Yet recall is a blunt and indirect response to a disputed registry act. It may occur after harm. It may replace one governor without correcting a record, preserving service or compensating a loss. It also depends on mobilisation across a defined membership, not the intensity of harm to a non-member.

Service Members do not possess the same election vote under the published bylaws. Customers of members may possess no ARIN corporate rights at all. The distribution is defensible as a membership design, but it prevents member democracy from serving as proof of public enforcement.

The distinction between voice and remedy matters here. A public consultation can let a customer submit evidence. A member discussion can expose concern. An election can change the future Board. None necessarily gives the customer a decision on whether the challenged act breached a duty or an order restoring its position.

Member rights may also be difficult to exercise. Petition percentages, quorum, election timing, information access and collective-action costs determine whether a paper mechanism can respond to a live controversy. Reporting only that a recall power exists tells little about accessibility, historical use or speed.

Corporate accountability should therefore be measured by outcomes and denominators. How many members were eligible? How many signed? Was information sufficient? How long did the process take? Did any later board action repair the underlying harm? A mechanism can be constitutionally real and operationally inadequate for a particular problem.

The same logic applies across RIRs. Member powers answer who controls the association. They do not answer whether an end user can challenge a service interruption, whether a regulator can enforce a statutory duty or whether recognition can be reviewed. Calling all of these “community accountability” conceals the missing link.

Contracts create standards while narrowing the claimant class

The ARIN Registration Services Agreement, version 14.0, dated 15 August 2025, replaces abstraction with defined promises. It identifies ARIN and the holder, describes services, incorporates policies to a stated extent, allocates obligations and establishes provisions for term, termination, assignment, liability and dispute handling.

This specificity makes legal review possible. A holder can argue that ARIN failed to provide a promised service, misapplied an incorporated policy or acted outside a stated termination ground. ARIN can point to the holder's representations and obligations. A court or other forum can read the words and evaluate evidence under governing law.

The same precision narrows access. The agreement states that it is for the parties' benefit and does not grant rights or remedies to other people or entities except where expressly provided. Customers are specifically contemplated in parts of the liability language, yet contemplation of harm does not make them parties.

The no-third-party-rights clause cannot necessarily eliminate mandatory law. It does show that the public-interest purpose and the service contract occupy different legal spaces. The charter may describe benefits for the general public while the agreement reserves contractual enforcement to identified parties.

This gap is where reputation can masquerade as accountability. ARIN can tell the public that it acts in the public interest, while telling a customer that only the holder has contractual rights. Both statements may be legally coherent. The governance question is whether another mechanism protects the customer when the holder cannot or will not act.

The answer might lie in the customer-provider contract, public law or a dedicated third-party review channel. Each has different consequences. A provider claim can correct the upstream relationship but may conflict with the provider's own interests. A statutory claim may address only particular conduct. A review channel may preserve operation without awarding compensation.

Accountability reporting should show these limits rather than presenting an agreement-based appeal as universally accessible. Who can file? What decision can be reviewed? What relief can be granted? Can a customer submit evidence or seek urgent preservation through the holder? Precision about exclusion is the first step toward filling it.

RIPE NCC demonstrates what a determinate duty looks like

The RIPE NCC Articles of Association, ripe-818, effective 1 July 2024, include provisions that contrast sharply with an abstract public-interest clause. They require the Executive Board to submit annual and financial reports to the General Meeting. After the relevant period, a member may commence legal proceedings against Executive Board members to enforce those specified obligations.

That provision has an identifiable claimant, respondent, duty and judicial route. It does not promise that every member will prevail, but it makes the obligation justiciable in a recognisable form. The standard is whether the specified reports were submitted as required, not whether the Board served humanity in the round.

The Articles also recognise arbitration for disputes between members and the Management Team over decisions concerning standard service agreements. The General Meeting has authority over appointment and dismissal of arbiters and changes to the procedure. Again, the subject and eligible parties are bounded.

The RIPE NCC Standard Service Agreement, ripe-812, dated November 2023, supplies more determinate terms concerning services, policies, liability and termination. The combination gives a member routes that can be matched to corporate reporting or service decisions.

It does not solve the public problem. A downstream user may not be a member. An arbitration about a member decision may not address region-wide competition, data protection or public reliance. The English publication of the Articles also notes that the Dutch text governs if translation differences arise, reinforcing the importance of jurisdiction and exact legal text.

The value of the comparison lies in design. A broad purpose can coexist with narrow, enforceable duties. Legislators, members and institutions need not make “public interest” itself the only standard. They can translate it into obligations to publish evidence, consider affected groups, give reasons, avoid conflicts, provide notice and permit independent review.

Specificity does not drain public purpose of moral force. It gives the promise points of attachment. A member can compel a report. An arbiter can review a service decision. A regulator can test a statutory standard. The public-interest clause then guides the system rather than carrying the whole weight of enforcement alone.

APNIC exposes the threshold problem

The APNIC By-laws trace APNIC's special-committee structure to a directors' resolution of 24 June 1998. They state community-serving entities, identify members as the governing body and allocate management to the Executive Council between general meetings.

They also provide a striking formal remedy. Review or amendment of an Executive Council decision requires an affirmative vote of two-thirds of the entire membership that was paid up at the specified time. This is more than a consultation. Members can alter a council outcome if they clear the threshold.

The denominator is decisive. Two-thirds of ballots cast is different from two-thirds of the entire eligible membership. The latter demands broad mobilisation and makes non-participation functionally supportive of the status quo. For a dispersed global membership, the power may be most useful for issues that generate exceptional engagement rather than urgent disputes affecting a small group.

The APNIC Membership Agreement offers more specific bilateral obligations. The company must maintain open communication mechanisms, consider member operational requests and provide rights and services under APNIC documents. A member can frame a claim around those terms more readily than around general benefit to the region.

Again, the wider public sits differently. APNIC's entities refer to service to the Internet community, training and public positions, but a customer of a member does not thereby acquire the member's vote or contractual rights. The public promise extends rhetorically farther than the corrective mechanism.

This does not prove the structure fails. Member elections, council accountability, open meetings and service arrangements can generate strong practical discipline. Reputation can influence an institution dependent on community cooperation. The conclusion is narrower: public-interest accountability cannot be measured merely by listing formal powers. Threshold, claimant class, timing and remedy determine usability.

A serious annual account would report the number of member requests, council reviews, attempted decision reversals, eligible voters, participation and resulting changes. Without outcome data, the public can see constitutional architecture but not whether it corrects contested power.

AFRINIC shows why broad entities require a legal map

The AFRINIC Bylaws describe a private company limited by guarantee under Mauritian law. Its entities include allocation and registration services, open and transparent communication, consensus-driven decision making, responsible resource management, public education and development of public policy in members' interests.

Those entities cover several constituencies. Some provisions focus on members. Others mention the African Internet community or public education. The income and capital are directed toward corporate entities rather than distribution to members, subject to stated exceptions. This is a meaningful nonprofit structure, not empty branding.

Yet breadth creates interpretive conflict. A decision might benefit current members while harming prospective entrants. Conservation may conflict with immediate access. Financial continuity may conflict with low fees. Security may conflict with procedural delay. Open participation may conflict with confidential evidence. The entities do not rank these interests.

The Board must make the initial balance. A reviewer then needs to know whether the question is corporate validity, directors' duties, a member right, a service contract, a statutory breach or recognition. Each route has a different plaintiff and remedy under Mauritian and other applicable law.

This is why international commentary should avoid treating all RIR purpose clauses as interchangeable. Similar words sit inside different corporate forms and jurisdictions. A Virginia nonstock corporation, Dutch association, Australian corporate committee and Mauritian company limited by guarantee do not derive authority from one global nonprofit code.

Comparative learning is still valuable. All can publish clearer standards, complaint outcomes, conflicts and remedial powers. All can state which constituency may use which mechanism. But a regional mission is not self-executing merely because another RIR uses similar language.

The legal map must also include public officials. Company registrars, tax bodies, courts, competition authorities, consumer regulators and data authorities may each have power over a slice of conduct. Their jurisdiction cannot be assumed, but neither should a private corporate complaint process be presented as a substitute for them.

Recognition is supervision only if someone can use it

The ICP-2 recognition criteria, accepted on 4 June 2001, require proposed RIRs to demonstrate community support, neutrality, technical capacity, financial stability, open policy development and service capability. All listed criteria are described as essential preconditions for recognition.

Recognition matters because a single RIR per large region is expected, making substitution difficult. The criteria give substance to the idea that an RIR holds a community-facing role rather than an ordinary commercial franchise. They can support evaluation of institutional legitimacy.

But recognition becomes accountability only if there is a current process for examining breach. Who may request review? Which body investigates? What evidence is available? Is the standard continuing compliance with recognition conditions, grave institutional failure or something else? Can recognition be conditioned, suspended or transferred? How is service continuity protected while the dispute proceeds?

Without answers, recognition is strongest at entry and weakest during operation. A corporation can point to historical approval as proof of legitimacy while affected people cannot use the recognition relationship to correct present conduct. The criteria then operate as pedigree rather than supervision.

Recognition is also too blunt for many disputes. Withdrawal or replacement of an RIR would have enormous operational consequences and cannot be the routine remedy for a mistaken record or unfair procedure. An effective regime needs intermediate tools: information requests, independent assessment, corrective plans, monitored milestones, preservation measures and transparent findings.

The distinction between institutional and individual relief must remain clear. A recognition review can improve governance without compensating a harmed holder. A contract appeal can correct a service decision without resolving region-wide legitimacy. A regulator can address competition or data conduct without deciding whether the RIR remains technically capable.

The public-interest promise needs these mechanisms to complement one another without pretending they form a single hierarchy. A complainant should be told which route addresses which harm and whether parallel proceedings affect one another.

The vanished final appeal is a warning about remedial gaps

RFC 7020, published in August 2013, describes the evolved Internet Numbers Registry System and its regional policy processes. It also records a consequential historical change. Earlier guidance had contemplated forwarding an appeal to IANA after other avenues were exhausted. RFC 7020 says RIR communities developed their own consensus-based appeal policies and that a further IANA appeal was no longer appropriate.

Decentralising appeal can be legitimate. Regional institutions understand their policies and legal settings. A global technical coordinator may be poorly placed to adjudicate member contracts or local law. Removing an unsuitable final level is not itself an accountability failure.

The warning lies in what replaces it. If each RIR's appeal is controlled by the same institution that made the decision, if access is limited to direct holders and if remedies cannot address downstream harm, the disappearance of final review leaves no equivalent check. Regional autonomy then becomes remedial finality.

RFC 7020 also expects broad participation reflecting functional, geographic and cultural diversity and consideration of affected parties. Those expectations support inclusive reasoning. As an informational technical document, however, it does not create a global tribunal or private claim.

The proper response is not to recreate IANA as a universal court. It is to specify the independence and scope of regional review, publish outcomes and reserve external legal supervision where applicable. Cross-RIR review standards could establish minimum safeguards without requiring identical corporate law.

The history also shows why appeal counts need context. A low number may reflect accurate first decisions, but it may also reflect narrow standing, cost, delay, fear of retaliation or a belief that the reviewer cannot change the outcome. Institutions should report both complaints received and reasons matters could not proceed.

An appeals system proves accountability through correction capacity, not mere existence. The critical evidence is whether it can preserve a contested position, access the decisive record, hear affected parties, publish reasons and require implementation.

Regulators can enforce law, not an undefined global good

Public regulators are often invoked as the answer to weak private remedies. An RIR is incorporated somewhere, uses contracts, processes data, employs staff, acquires services and affects markets. It is not outside law. Depending on jurisdiction and conduct, company, tax, competition, consumer, data, insolvency and other authorities may act.

Regulatory availability should not be overstated. A company registrar may enforce filing or governance rules without reviewing number policy. A tax authority may examine exempt-purpose compliance without correcting a resource transfer. A competition authority may require evidence of market power and exclusionary conduct. A data authority may address processing but not registry ownership. Each regulator has a statutory mandate.

The public-interest clause may be relevant evidence within those mandates. It can show declared purpose, expected beneficiaries or inconsistency in the institution's reasons. It does not expand a regulator's jurisdiction beyond law. Nor does it guarantee that a regulator will prioritise a technically specialised dispute.

Cross-border operation adds complexity. The RIR's incorporation jurisdiction may differ from the customer's location, the holder's domicile and the place where operational harm appears. Contract forum clauses may govern some claims but not mandatory regulation. Recognition operates at another institutional level. A complainant needs a route map, not a generic direction to contact authorities.

RIRs should publish that map without offering legal conclusions for every case. It can identify the incorporation body, governing contract forum, independent service review, data contact, competition jurisdictions relevant to major operations and emergency continuity contact. It should also state which complaints the RIR's own bodies cannot decide.

Regulators, in turn, should preserve functional boundaries while coordinating. A finding about corporate governance may trigger a service review. Evidence of systematic denial may interest competition authorities. A data correction can affect registry accuracy. Coordination is useful; jurisdictional vagueness is not.

Public law is therefore a backstop made of specific powers, not a single guardian of the Internet's public interest. The same enforceability test still applies: eligible complainant, defined standard, competent forum and available remedy.

A usable public-interest standard must expose the balance

No drafting can eliminate judgment from public-interest decisions. The Internet serves diverse users, and registry choices involve security, scarcity, cost, stability, privacy, competition and administrative feasibility. A rigid formula would create false precision.

The standard can nevertheless require disciplined balancing. First, legality: the institution must act within charter, bylaws, agreement, policy and applicable law. Second, evidence: decisive factual claims must be identified and open to challenge where confidentiality permits. Third, affected interests: members, non-members, customers and system operators must be mapped rather than merged into “the community.”

Fourth, equal treatment: materially similar cases should receive similar treatment, with reasons for departure. Fifth, necessity: the institution should connect the measure to a legitimate registry objective. Sixth, proportionality: the expected benefit should be weighed against foreseeable harm and less restrictive alternatives. Seventh, continuity: critical services and innocent downstream users should receive protection where consistent with security.

Eighth, conflicts: decision-makers should disclose interests and recuse where required. Ninth, reasons: the final decision should explain how evidence and competing interests were treated. Tenth, review: an affected party should know who can reconsider the decision and what that reviewer may order.

These requirements do not dictate one substantive result. They make the balance inspectable. A registry can still act urgently against fraud or security risk, but it must explain the emergency basis and submit to prompt confirmation. It can charge sustainable fees, but should publish the cost and distribution analysis. It can deny a transfer, but must identify the rule, evidence and review route.

A court or panel applying this standard would not need to decide the best Internet policy from scratch. It could test whether the institution remained within authority, used reliable evidence, considered mandatory interests and reached a proportionate outcome. The remedy might be reconsideration rather than substitution.

This translation from purpose to procedure is the central design task. Broad public-interest language becomes enforceable through narrower duties that reviewers can observe. Without those duties, the clause remains infinitely adaptable to whatever the institution already decided.

Complaint statistics must measure correction, not contact

Institutions often report the number of inquiries or consultations as evidence of accountability. Contact volume measures accessibility, not remedy. A person can receive a courteous response while the challenged decision remains untouched and the forum lacks authority to change it.

A credible public register should classify complaints by claimant status: voting member, service member, contracted holder, legacy holder, sponsored party, downstream customer, network operator, public body or other affected person. It should identify the challenged function, such as membership, allocation, transfer, registration data, reverse DNS, routing security, fees, elections or disclosure.

The register should then show the asserted standard and forum. Was the matter treated as a policy appeal, service dispute, corporate complaint, ethics issue, data request, contractual claim or regulatory referral? Was it dismissed for standing, lateness, jurisdiction or limited public evidence evidence? Those outcomes reveal whether public beneficiaries can enter the system.

Remedy data are essential. Did the body preserve the status quo, correct a record, restore service, order a new decision, change a policy, publish reasons, remove an officeholder, refer a breach or provide no relief? How long did interim and final action take? Was the result implemented?

Aggregate publication can protect confidential and security-sensitive material. Redacted reasoned decisions can explain standards without exposing private data. The goal is not spectacle but institutional learning and public verification.

Denominators prevent misleading success rates. Ten corrected complaints mean something different among twelve completed matters than among ten thousand rejected contacts. Participation rates need the eligible population. Appeal rates need the number of appealable decisions. Timeliness needs the distribution, not only an average.

This evidence would also distinguish reputation from accountability. If complaints generate reasoned correction, the promise operates institutionally. If they produce acknowledgements but no body can change an outcome, the promise mainly protects confidence. Both effects can matter, but they should not share the same label.

The remedy must fit the failure

Public-interest disputes range from a missing notice to systemic capture. One universal sanction would be either trivial or destructive. The remedial architecture needs a ladder whose rungs correspond to the function and urgency.

For a contested service act, the first need may be preservation. A short hold on deregistration, certification change or transfer can prevent irreversible harm while evidence is tested. Preservation should require a threshold showing and security safeguards; it should not become a tool for indefinite delay.

For factual or procedural error, correction or a new decision may be enough. The reviewer should identify the defect, require a lawful process and set a deadline. Substituting its own policy judgment should be reserved for cases where the rules permit it and the record is complete.

For governance failure, remedies may include disclosure, recusal, a member vote, director removal, amendment of procedures or independent monitoring. Those measures protect the institution but may not repair an individual loss. Contractual or legal relief may still be needed.

For systematic statutory breach, referral to a competent regulator or court may be appropriate. Recognition review may address persistent institutional inability or refusal, but continuity planning must precede any structural remedy. The public should not lose registry service in the name of protecting it.

Compensation is a separate question. Agreements may limit liability; public law may constrain those limits; causation across provider and customer relationships can be complex. A complaint body that cannot award damages should say so and preserve access to other forums rather than imply complete resolution.

The remedy should also be enforceable against management. A recommendation that can be declined without reasons is advice. If a Board may reject it, the rejection should be public, reasoned and subject to a higher governance consequence. If an order is binding, the source and limit of that power should be clear.

Accountability is not maximal punishment. It is reliable correction. The test is whether the system can move the affected legal or operational position toward compliance before harm becomes irreversible.

A design for promises that can be used

The first reform is a public-interest schedule attached to each RIR's constitutional documents. It should translate mission language into reviewable duties: identify affected groups, disclose conflicts, use evidence, give reasons, consider less harmful alternatives, protect continuity and provide review. The schedule can preserve discretion while making its exercise observable.

Second, every review channel should publish a jurisdiction card. It should state who may file, which decisions are covered, the filing period, available interim measures, standard of review, access to evidence, reviewer appointment, possible remedies and implementation rules. A general complaints address is not enough.

Third, non-members need a route to present direct evidence of harm. They need not receive corporate votes or automatic damages. They should be able to trigger triage, seek urgent preservation where thresholds are met and participate in a holder's review when their interests cannot otherwise be represented.

Fourth, review should be structurally separate from the people who made or advised on the first decision. Appointment by members or the Board can be compatible with independence if terms, conflicts, removal and budget are protected. Published reasoned decisions create a body of standards and reduce selective treatment.

Fifth, corporate, contractual, recognition and regulatory channels should exchange referrals without merging authority. A service panel should be able to flag a governance pattern. A Board committee should refer statutory concerns. A recognition body should receive evidence of persistent failure. Each remains responsible for its own legal test.

Sixth, institutions should publish the complaint and remedy dataset described above. Repeated findings should trigger policy review, training or governance action. Absence of complaints should prompt accessibility checks rather than immediate self-congratulation.

Finally, boards should state when a public-interest claim is not justiciable in a given channel. Candour about limits is better than ceremonial reception. The response should direct the complainant to the nearest competent body and preserve deadlines where possible.

These reforms do not constitutionalise every disagreement. They make existing power answerable. A claimant may still lose because the decision was lawful, evidenced and proportionate. The difference is that the public-interest promise would have been applied rather than advertised.

Evidence limits and the test ahead

Public RIR documents show purposes, membership structures, agreements and selected review mechanisms. They do not provide a complete cross-regional denominator of complaints, standing decisions, interim protections, corrections and implementation times. Without that evidence, no confident claim can be made that public-interest promises are either universally effective or universally empty.

Legal effect also varies. ARIN's charter sits in Virginia law. RIPE NCC's association in Dutch law, APNIC's structure in Australian law and AFRINIC's company in Mauritian law create different routes for members, directors, officials and outsiders. Contract terms and mandatory law further divide cases. The analysis here is institutional, not a substitute for jurisdiction-specific advice.

Reputation should not be dismissed. RIRs depend on broad voluntary coordination, technical trust and community participation. A well-documented criticism can influence elections, policy and recognition even without a court order. Soft pressure can be materially effective. The error is calling it an enforceable remedy when no decision-maker must respond.

The next test is empirical. For each major complaint, ask who filed, what standard was invoked, who decided, what evidence was available, what relief could be granted and whether the outcome changed. Compare members with non-members. Compare routine correction with urgent service preservation. Publish failures as well as successes.

The phrase “public interest” should survive that test. Internet number coordination genuinely affects people far beyond an RIR's members, and corporate governors should be required to consider them. But the phrase must be attached to duties that can be observed and remedies that can be used.

Without plaintiff, standard, forum and remedy, the clause is a reputation tool. With them, it can become a disciplined promise: not that every complainant wins, but that consequential registry power can be tested, corrected and explained to the public in whose name it is exercised.