Summary

  • ICANN has no statutory members, and its bylaws expressly say that the Empowered Community is not an ICANN member. The Empowered Community is instead a separate California nonprofit association made up of five Decisional Entities and recognised as ICANN's sole designator of directors.
  • The five Decisional Entities are the ASO, ccNSO, GNSO, ALAC and GAC. Their representatives form the Empowered Community Administration, but those representatives may transmit notices and tally decisions only as directed by their institutions; they are not an independent upper chamber.
  • An individual can start some procedures only by petitioning a Decisional Entity under that entity's own rules. Acceptance by one institution is merely the first gate. Most contested actions must then obtain support from at least one other Decisional Entity before a community forum and final vote can occur.
  • Final thresholds are intentionally high. Most rejection actions and a recall of the whole board require support from four Decisional Entities and no more than one objection. A standard bylaw rejection usually requires three supporters; a Community IRP usually requires three. Other powers have distinct thresholds and safeguards.
  • The public record supplies unusually clear practical tests. ALAC's June 2025 petition seeking community reconsideration of the postponement of the fourth Accountability and Transparency Review ended when no other Decisional Entity supported it. ALAC's June 2026 petition to reject a standard bylaw amendment also ended without a second supporter.
  • These failures do not show that the rights are fictional. The same machinery approved Fundamental Bylaw changes in 2023 and July 2026, regularly designates board members and records the expiry of unchallenged budget and bylaw notices. It is effective at formal consent and certification when institutional positions converge.
  • The architecture is better suited to deterring or correcting an extreme, legible board departure than to answering routine grievances. It requires cross-constituency agreement, protects stability and limits capture by one sector, but it also converts diffuse public concern into five institutional veto points.
  • Accountability should therefore be judged at each link: who may petition, how a Decisional Entity decides, whether it gives reasons, which other entity supplies support, what final threshold applies and what remedy follows. Counting only formal powers exaggerates access; counting only successful sanctions understates deterrence.

A constitutional objection reaches its second gate

On 1 June 2026, the At-Large Advisory Committee did something the post-transition ICANN constitution was designed to permit. Acting as one of the five Decisional Entities, ALAC delivered a petition notice seeking rejection of a standard bylaw amendment. The amendment concerned the timing of ICANN's Specific Reviews. ALAC argued that it would formalise long delays in mechanisms intended to examine accountability, security and registration-directory policy. This was not a comment sent into a general consultation. It was a notice invoking a defined legal power.

The distinction soon became decisive. Under Annex D of the ICANN Bylaws, one Decisional Entity can accept a timely petition. It cannot by itself carry an ordinary rejection case to a community forum. The petitioning body has seven days after the petition period to obtain support from at least one other Decisional Entity. ALAC invited all four peers to support its case. The ccNSO Council decided not to do so after considering comments from country-code managers. The ASO reported that its regional Internet registry communities had supplied no input supporting the petition and therefore also declined. No other supporter arrived before the deadline.

On 9 June, the Empowered Community Administration certified termination. The amendment survived without a community forum and without the final vote that would have required three Decisional Entities to support rejection. The cause was not a judgment by ICANN's board, its legal staff or an external tribunal. It was failure at the coalition gate built into the community mechanism.

That episode offers a more useful description of the Empowered Community than a list of imposing rights. A recognised constituent body identified a bylaw-level grievance, used its own authority to accept it, published detailed grounds and asked peers for support. The system made the challenge visible and time-bounded. It also stopped the challenge before its merits were tested collectively. Power existed at every stage, but no single stage belonged to “the community” in the undifferentiated sense.

ICANN has no members, by deliberate design

The title of the institution contains its central paradox. Article 23 of the bylaws says that ICANN shall not have members within the meaning of the California Corporations Code and adds, for avoidance of doubt, that the Empowered Community is not a member of ICANN. Article 6 nevertheless creates the Empowered Community as a California nonprofit association. Its five constituent organisations are called associates and, where interpretation of the association's governing rules requires it, may be treated as members of that association.

There are therefore two corporate relationships, not one. ICANN is a California nonprofit public-benefit corporation without statutory members. The Empowered Community is a separate association with five associates. It has legal personhood for its stated purposes, but it does not sit above ICANN as an owner in the ordinary commercial sense. It may not acquire property, employ staff, appoint its own officers or pursue activities beyond the rights listed in ICANN's articles and bylaws.

This structure was a response to a hard problem in the 2016 transition. The end of the United States government's IANA functions contract removed an external counterparty with a renewal decision and a recognisable contracting officer. The transition designers wanted enforceable community powers without turning ICANN into a mass-membership corporation, exposing every entity to uncertain statutory duties or pretending that the world's Internet users could be entered on a corporate register.

The CCWG-Accountability proposal settled on a “sole designator” model. A designator under California law can select and remove directors without possessing the full bundle of rights attached to a statutory corporate member. The proposal paired that status with specially protected bylaws, escalation procedures and standing to enforce decisions. It was a constitutional engineering choice: give selected community institutions legal handles on the corporation while declining to create a universal franchise.

Calling the result membership accountability is thus analytically useful only with a qualification. It substitutes specified designator and consent rights for membership. It does not reproduce annual member meetings, a direct membership roll, equal votes among natural persons or a general residual right to direct the corporation.

The sole designator is powerful but narrow

Since 1 October 2016, the Empowered Community has been ICANN's sole designator. It formally designates all directors other than the president, who serves ex officio. In practice the nominations come from the bodies allocated particular seats: the Nominating Committee, the supporting organisations and the At-Large community. The Empowered Community Administration records and transmits the resulting designations. The designator status is the legal bridge that turns those selection decisions into board seats.

The same bridge supports removal. The bylaws empower the community to remove individual directors other than the president and to recall the full board other than the president. They also allow the community to reject ICANN and IANA budgets, operating and strategic plans, standard bylaw changes and specified decisions concerning the post-transition IANA arrangements. Fundamental Bylaw amendments, amendments to the articles and a sale of substantially all assets require community approval.

The community may initiate reconsideration, mediation or an Independent Review Process, and it may pursue enforcement in a court with jurisdiction over ICANN.

Those are not advisory privileges. A valid rejection of a standard bylaw amendment makes it null. Failure to obtain required community approval makes a proposed Fundamental Bylaw amendment ineffective despite board approval. A valid removal notice ends a director's tenure. Article 6 recognises the association's ability to sue and prevents ICANN from denying its legal personhood as a defence.

But the powers are enumerated. Article 6 says the association's sole purpose is to exercise the rights assigned to it and that an act outside the prescribed bylaws has no effect. It cannot write ordinary policy, manage ICANN's staff, negotiate registry contracts at will or take over IANA operations. Even in constitutional matters, the community cannot directly propose a bylaw amendment. It may reject a standard amendment or withhold approval from a fundamental one proposed through the board procedure; it does not become a parallel legislature with a general initiative power.

This combination matters for legitimacy. A body with narrowly stated legal powers can be an effective circuit breaker without becoming an unaccountable alternative government. Yet narrowness also means that a complaint must be translated into one of the recognised actions. A concern about culture, staffing, delay or implementation may be politically important without fitting a budget rejection, director removal, reconsideration request or review claim.

Five institutions stand between an individual and the legal act

The Empowered Community consists of the Address Supporting Organization, the Country Code Names Supporting Organization, the Generic Names Supporting Organization, the At-Large Advisory Committee and the Governmental Advisory Committee. Each brings a different institutional constituency. The ASO links to the regional Internet registry system and number-resource policy. The ccNSO organises participating country-code top-level-domain managers. The GNSO contains contracted and non-contracted interests concerned with generic names. ALAC is the formal channel for individual Internet-user interests within ICANN.

The GAC brings governments and public authorities.

This is pluralism by institutional type, not apportionment by population. The five are not equal-sized electorates and do not claim equivalent membership forms. A GNSO Council vote may reflect houses and stakeholder groups. A ccNSO decision begins among ccTLD managers that have joined the ccNSO. An ASO position may be formed through the Number Resource Organization and regional communities. ALAC draws on regional At-Large organisations and accredited structures. GAC decision-making follows governmental committee rules.

Their one institutional position each at the final community stage does not imply that the people or entities behind them are numerically equal.

The bylaws leave an important part of access to each entity. Section 6.1(g) requires every Decisional Entity to adopt rules explaining who may submit a petition, how a petition is filed, how the institution accepts or rejects it, how it decides whether a matter is resolved, how it supports or entities to another entity's action and how it notifies constituents. Annex D then places the common outer procedure around those different internal rules.

An individual therefore has no universal button marked “Empowered Community”. For certain actions, an individual may petition a Decisional Entity, subject to that entity's requirements. The institution decides whether to accept. The petitioner may need to belong to, participate in or persuade a particular constituency. A person who can readily approach ALAC may have no comparable standing in the GNSO or ccNSO. Conversely, a registry operator or number-policy entity may have established channels unavailable to a sporadic end user.

The representation chain is a feature and a filter. It ensures that a drastic corporate remedy is not launched by an unauthenticated message. It also means access depends on internal constitutional literacy, meeting calendars, language capacity, volunteer time and the willingness of institutional officers to treat the complaint as their concern.

The Administration administers; it does not originate a mandate

The five representatives serving in the Empowered Community Administration are easy to mistake for five governors. The bylaws give them a more constrained role. Each representative must act solely as directed by the represented Decisional Entity and under that entity's procedures. Collectively they send and receive notices, organise or moderate community forums, tally decisions and implement valid outcomes. ICANN may rely on their notices as evidence that the relevant institution or the association acted properly.

The official Administration page describes the body in similarly practical terms: receiving and sending notifications, moderating calls and forums, and tallying entity decisions. In mediation it appoints representatives and helps select potential mediators. It has no separate electorate and no authority to substitute its collective judgment for instructions from the five institutions.

That separation reduces a familiar agency risk. A representative cannot arrive in the Administration and bargain away a constituency's position merely because the representative finds compromise convenient. The authenticated notice is backed by an institutional decision. It also limits speed. When a short deadline runs, each representative may need to return through council meetings, consultations, ballots or other internal steps before providing a valid answer.

The Administration's neutrality is most visible when a case ends. In June 2026 its termination letter was signed in the name of the Administration, including the ALAC representative whose institution had brought the petition. The letter did not decide that ALAC's accountability concerns lacked merit. It certified the procedural fact that no other Decisional Entity had provided support by the deadline. This is valuable institutional hygiene. The recorder does not become the judge.

It also explains why assessing the Administration's activism misses the point. The relevant questions are whether notices are accurate and timely, forums are fairly managed, decisions are tallied correctly and constituent instructions are authentic. Political initiative sits earlier, inside the Decisional Entities.

Rejection is a relay, not a vote taken at once

For budgets, plans, standard bylaws and specified IANA-related decisions, the rejection procedure has several stages. First, the board or secretary sends formal notice. For twenty-one days, an individual may submit a petition to a Decisional Entity under that entity's rules. The entity must accept or reject it. Acceptance produces a public notice with reasons; for budgets and plans, the stated issue generally must have been raised during the relevant public comment period.

Second, the petitioning entity seeks one supporting entity. The support period is short: it ends seven days after the first petition period. Without a second institution, the case terminates. With support, ICANN convenes a community forum at the Administration's direction. The forum can include a public conference call, written submissions, board and staff participation and remote attendance. It is expressly not a decisional body.

Third, after the forum period, each Decisional Entity has twenty-one days to record support, objection or abstention. Silence counts as abstention. For most rejection actions, four of five entities must support and no more than one may entity. A standard bylaw rejection generally requires three supporters, including the policy-development body in specified circumstances, and no more than one objection. The GAC carve-out can adjust participation and thresholds where the challenged board resolution implements GAC consensus advice.

These stages perform different functions. The first tests whether one recognised constituency will own the complaint. The second tests whether the issue crosses an institutional boundary. The forum tests reasons publicly and permits resolution. The final stage tests broad agreement for the remedy. Combining them under one word, “vote”, conceals where accountability succeeds or fails.

The relay also changes incentives. A petitioner must anticipate not just the accepting institution's concerns but those of a potential supporter. A Decisional Entity may agree that a problem exists while rejecting the requested remedy. A body may prefer negotiation, a future review or an ordinary reconsideration request to the constitutional cost of nullifying a budget or bylaw. The absence of support can therefore mean disagreement about facts, remedy, timing, institutional competence or priority. A termination notice alone cannot reveal which.

Approval has a different constitutional logic

Fundamental Bylaw amendments, amendments to ICANN's articles and a sale of substantially all assets are Approval Actions. They begin with a board proposal rather than an adverse petition. A community forum follows, and each Decisional Entity records support, objection or abstention. Fundamental Bylaw and articles changes ordinarily require support from at least three entities and no more than one objection; an asset sale requires at least four supporters and no more than one objection.

The asymmetry is sensible. A fundamental change should not take effect merely because nobody files a timely objection. It needs affirmative community consent. By contrast, an annual budget cannot wait indefinitely for affirmative approval from every constituency. It takes effect unless the rejection machinery reaches its threshold. The constitution distinguishes protected change from recurring administration.

The public record shows this affirmative power working. In 2023 the board proposed amendments concerning the IANA Naming Function Review and separation arrangements. The Empowered Community correspondence archive for 2023 records a community forum, support notices from the Decisional Entities and the Administration's approval notice. In July 2026, the Administration approved further Fundamental Bylaw amendments concerning the Customer Standing Committee and the IANA Naming Function Review after four entities supported and the GAC abstained.

These are not dramatic confrontations, but they are genuine exercises of constitutional power. Without the required approval, the amendments would have failed despite board action. Their consensual character should not make them invisible. Constitutions often do their most consequential work by requiring proponents to assemble consent before change occurs.

At the same time, approval cases are easier for the mechanism than contested sanctions. The board has usually developed the proposal through consultation, and entity institutions may already have helped shape it. The procedure confirms convergence. A rejection petition starts from conflict and asks peers to oppose a completed board act. Comparing raw counts of approvals and sanctions would therefore say little about equal accessibility or courage.

Director removal is not one uniform power

The statement that the community can remove directors hides three distinct procedures. Directors selected by the Nominating Committee may be challenged through a petition accepted by a Decisional Entity, supported by another entity, discussed in a community forum and decided under the Annex D threshold. Directors associated with a Supporting Organization or At-Large seat are handled primarily by the applicable appointing entity. A whole-board recall requires an even broader coalition.

For a director appointed through a Supporting Organization or ALAC, an individual may petition the relevant Decisional Entity. The institution must invite the director and board chair to dialogue before accepting. After a public comment period and community forum, removal requires a three-quarters majority within the applicable entity under its internal rules. If that threshold is not achieved, the same director cannot be subjected to another such procedure for the remainder of the term under the specified condition. The design protects against repetitive harassment while preserving appointing-body control.

For a Nominating Committee director, a petitioning entity needs at least one supporting entity to proceed. The final removal decision has its own cross-community threshold. A recall of all directors other than the president must begin with one entity and obtain support from at least two others merely to reach the forum. Final recall requires four supporters and no more than one objection, subject to the GAC rule in the defined circumstances.

These safeguards reflect the severity of the remedy. Removing a director is not equivalent to rejecting a line item. A recall can destabilise oversight, trigger simultaneous vacancies and create incentives for a board to cater to short-term constituency pressure. Dialogue, public reasons and high thresholds make the power credible as a last resort rather than routine parliamentary no-confidence.

They also dilute direct electoral analogy. The public does not vote to remove a director. Even the institution that helped select a director must follow a special procedure, and directors owe duties to ICANN rather than serving as instructed delegates of their selectors. The community power is a constitutional remedy for loss of confidence, not a mandate system in which each director transmits a constituency's daily preferences.

The first practical test came through reconsideration

The 2026 rejection attempt was preceded by a closely related test. On 19 June 2025 ALAC filed a Community Reconsideration Petition challenging board decisions that continued the deferral of the fourth Accountability and Transparency Review and directed work on a redesigned review programme. ALAC argued that the five-year timing in the bylaws had been exceeded and that a community-led review function was being displaced.

A Community Reconsideration Request is not simply an ordinary request filed by an affected person. When brought by the Empowered Community, one Decisional Entity initiates the petition and then needs support from at least one other entity within twenty-one days. If support arrives, a forum and final community decision follow. The eventual request would ask ICANN's reconsideration machinery to review board or staff action; it would not itself reverse the action on filing.

The second supporter did not arrive. The ccNSO publicly explained that it shared concern about review delays but believed the review system needed a pause and reassessment before another review began. Its disagreement was thus about remedy and sequencing, not a denial that accountability had a problem. On 11 July 2025 the Administration certified that no other Decisional Entity had supported ALAC before the deadline and terminated the initiation procedure.

This case reveals both the virtue and the weakness of the threshold. One institutional constituency could not relabel its preferred answer as the view of the global community. Another could acknowledge the same institutional pathology while refusing the proposed cure. Requiring a coalition prevented one sector from monopolising the community's legal standing.

But the case also exposed the distance between a bylaw promise and an available remedy. If a concern is about delayed accountability review, waiting for five institutions with different workloads and incentives to agree on how to restore review can reproduce the delay. The mechanism filters for shared urgency as well as substantive merit. Ordinary accountability can fail because the issue is not simultaneously urgent to two institutional agendas.

The 2026 petition made the representation chain observable

The following year's standard-bylaw petition sharpened the lesson. ALAC did not merely repeat its 2025 request. It challenged a board-approved amendment that would insert transition provisions on Specific Reviews into the bylaws. Because a standard bylaw amendment is expressly a Rejection Action, the complaint fitted a more direct constitutional category. ALAC's notice identified the board action, cited the bylaw authority, described public-comment concerns and requested rejection.

The ASO's response shows how a global-sounding power is translated regionally. The NRO Executive Council sought input from the regional Internet registry communities on whether the ASO should support. It reported receiving no supportive input and declined. That does not establish how every address holder or network operator viewed the amendment. It establishes the result of the ASO's stated consultation route for that decision.

The ccNSO held a special council meeting, considered comments from the country-code community and declined support. Again, the decision was not a poll of all country-code registrants or all national Internet users. It was a decision by the recognised Decisional Entity under its governance arrangements. GNSO and GAC support did not arrive before expiry. The Administration's letter accurately reduced the outcome to the fact required by Annex D: no other entity supported in time.

The action therefore passed a more demanding test of visibility than many international governance disputes. The initiator, grounds, target, peer invitations, two refusals and termination time are public. Researchers can identify the exact gate at which the case stopped. What remains less complete is comparative reasoning. The record does not provide one standard decision template across all five entities, a common disclosure of consultation participation or a consolidated explanation from every non-supporter.

The failure was neither total voicelessness nor full adjudication. ALAC could trigger the first legal stage. It could not trigger the association as a whole. That is the precise answer to who can activate an Empowered Community right: initiation is distributed, while exercise is collective.

Routine notices show a machine built to wait

Most entries in the Empowered Community correspondence archive are less contentious. The secretary gives notice of an approved ICANN budget, IANA budget, operating plan, strategic plan or standard bylaw amendment. The petition period passes. The Administration certifies expiry of the rejection opportunity. Board-seat selectors send nominations; the Administration confirms designations. Representatives are appointed or replaced.

It would be wrong to describe every expiry letter as a failed challenge. In many cases nobody sought rejection because consultations had resolved objections, because the action was acceptable, because institutions assigned attention elsewhere or because potential opponents judged the remedy disproportionate. The archive does not supply the counterfactual. Nor should routine board designations be mistaken for independent selection by the Administration; they usually give legal effect to a choice made by the assigned nominating body.

Nevertheless, routine administration matters. A constitutional right with no notice, clock or authenticated record is difficult to use. The repeated letters prove that budget and bylaw decisions enter a known challenge window and that somebody certifies when it closes. They also create an evidence trail for a court or review panel if the board attempted to implement an action before the community period expired.

The machinery is designed to wait. That waiting imposes a modest constraint even when no petition appears. A board knows that certain actions cannot become unassailable immediately. Staff must produce notice, preserve the approved text and allow institutions to consult. For Fundamental Bylaws, waiting is stronger: the amendment remains ineffective until affirmative approval.

This is deterrence through procedure, not evidence that every board choice is substantively accountable. The open question is whether the waiting period gives constituencies enough practical time to recognise a problem, mobilise expertise and complete their internal decisions. Twenty-one days can be long for a corporate filing and short for a volunteer-led global association spanning languages and meeting cycles.

High thresholds protect against capture and protect the status quo

The case for demanding four of five institutions to reject most major actions is strong. ICANN coordinates unique identifiers used across jurisdictions and markets. A budget failure or board recall can create operational uncertainty beyond the faction that initiated it. Cross-community support shows that the objection is not confined to one commercial segment, one technical community or one set of governments. The limit of no more than one objection prevents a bare coalition from acting over deep resistance from multiple sectors.

High thresholds also answer a capture problem. A well-organised interest might dominate one internal council or mobilise one petition channel. It is harder to capture institutions built on different constituencies at the same time. The need to explain a case across institutional vocabularies can improve reasons: a names-policy complaint must become intelligible to address-policy, country-code, user and governmental entities.

Yet every supermajority rule favours the existing decision. The board needs to act under its own voting rules once; opponents must assemble a dispersed coalition under a clock. Abstention does not help a petition, even when it reflects uncertainty rather than approval. Institutions with no direct stake may rationally avoid the cost of taking sides. A proposal can therefore survive despite concern in several constituencies if concern is not converted into affirmative support for the specified remedy.

This status-quo bias is most defensible for catastrophic remedies. It is less obviously desirable at the early support gate. Requiring one additional institution before a forum keeps frivolous cases from consuming resources, but it can also prevent the public exchange that would allow uncertain institutions to learn. In 2025 and 2026 the cases ended before a community forum. The decision architecture demanded preliminary coalition before collective deliberation.

Changing the final thresholds would risk making constitutional power too easy to wield. Improving the early stage need not. A non-decisional hearing, a standard written response from every entity or a longer support period for complex amendments could expose reasons without lowering the vote needed to overturn board action.

Governments are inside the coalition, not above it

The Governmental Advisory Committee is one Decisional Entity. That position is constitutionally significant because the 2016 transition was expressly designed not to replace United States stewardship with control by governments or an intergovernmental body. GAC participation gives public authorities a voice inside the community association; it does not give them unilateral command.

There is also a specific conflict rule. When a challenged board resolution implements GAC consensus advice and that advice was a material factor, the GAC cannot participate as a decision-maker in the community challenge to that implementation. Thresholds adjust accordingly. The principle is straightforward: an institution should not supply both the advice driving the board action and a vote deciding the community challenge to that same action.

The carve-out is narrower than a general exclusion of governments. It depends on identified GAC consensus advice and the board's stated reliance. GAC may participate in other community matters, including approval actions. In July 2026 it abstained on the Fundamental Bylaw amendments while four other entities supported them; the amendments still met the required threshold.

This arrangement distributes public authority but does not dissolve it. Governments retain their domestic powers outside ICANN. Inside the Empowered Community, they occupy one institutional channel subject to the association's rules. The model thereby avoids two extremes: pretending governments have no legitimate interest in the root and unique identifiers, and allowing governmental consensus to become an unreviewable source of both policy and constitutional enforcement.

For ordinary users, however, the addition of government as a fifth institutional constituency does not create direct representation. A person may petition ALAC, lobby a national administration, participate in a GNSO constituency or engage another body, but those routes have different access conditions. Institutional diversity is not the same as an equal franchise.

Enforceability begins where consensus ends

The community's strongest claim to be more than advisory lies in the remedy after a valid decision. Article 6 allows it to enforce rights through Annex D and in a court with jurisdiction over ICANN. ICANN recognises the association's legal personhood and agrees not to challenge standing on that basis. For director removal and board recall, ICANN cannot insist that reconsideration or independent review had to be completed first, subject to the specified exception.

The Independent Review Process supplies another legalised route. A Community IRP petition can be initiated by a Decisional Entity, but it needs a second entity for the petition to proceed and generally three supporters at the final initiation decision, with no more than one objection. The Empowered Community is deemed materially affected by covered actions and receives protections against standing objections. If it prevails, the result can establish that board or staff action violated ICANN's articles or bylaws.

These provisions answer a defect common in voluntary governance: the body accused of violating the rule cannot end the matter simply by refusing to listen. A recognised claimant, stated standard and external panel or court make defiance costlier. Fundamental provisions also protect the enforcement structure itself from unilateral board amendment.

But enforceability does not solve activation. A court right is irrelevant if no Decisional Entity accepts a petition, if no second entity supports it or if the final coalition fails. Nor does legal standing guarantee a quick merits decision, affordable representation or a remedy tailored to institutional delay. The architecture is strongest after agreement has formed. Its most uncertain point is forming that agreement among communities with unlike interests.

This helps explain the thesis. The Empowered Community can constrain extreme board conduct because an unmistakable departure may align several institutions and unlock binding remedies. It is weaker at ordinary accountability because ambiguity, distributed costs and different preferred cures prevent the coalition from forming.

Representation must be audited one link at a time

It is tempting to ask whether the Empowered Community represents “the Internet community” and answer yes or no. Neither answer is sufficiently precise. Representation differs by Decisional Entity, issue and decision stage. A useful audit starts with the person or organisation affected and follows the claim through every institutional handoff.

First, which entity's rules allow the claimant to petition? Second, what information, membership or endorsement is required? Third, which council, committee, house or community consultation decides acceptance? Fourth, is the vote and rationale public? Fifth, how does the entity seek a second supporter? Sixth, do peers explain support, opposition or silence? Seventh, does the final community threshold match the severity of the remedy? Eighth, can the resulting decision be enforced without returning discretion to the same board it constrains?

The 2026 record performs well on several points. ALAC's grounds, peer invitation, ASO and ccNSO answers, deadline and termination are public. It performs less well on comparability: the public cannot read one consolidated account of how each entity consulted, how many eligible voices engaged, what alternatives they considered and why two institutions did not provide support notices. No responsible conclusion about a global support rate can be drawn from those missing denominators.

The audit should also distinguish institutional authorisation from sociological representation. ALAC is authorised to speak through the At-Large structure, but that does not mean every Internet user participated. The ASO can state a position formed through its accepted number-community procedure without claiming every address holder agreed. The GAC can record governmental consensus without turning it into a vote of every resident. Legal validity and representative depth are related but separate.

Publishing those limits would strengthen rather than weaken legitimacy. A body can say exactly whom it consulted, under which rule and with what participation, while avoiding the fiction that five notices directly aggregate humanity.

A deterrent cannot be measured only by successful explosions

The low visible incidence of board recalls, director removals or budget rejections could support opposite stories. One says the mechanism is ceremonial because its thresholds prevent use. The other says the credible threat induces the board to consult and compromise before formal action. The public record does not permit either story to be universalised.

Deterrence is difficult to observe because the relevant events do not happen. A board may revise a proposal after hearing that one Decisional Entity could petition. Staff may address a concern during public comment. Entity leaders may decide privately that a concession removes the need for escalation. Conversely, institutions may decline to challenge because coordination is costly, not because the board was responsive.

Better evidence would record near-use without exposing privileged negotiations. Board notices could be linked to material concerns raised by Decisional Entities, changes made before approval and any stated intention to consider community powers. Entities could publish annual accounts of petitions received, accepted, rejected or withdrawn, with reasons and privacy protections. The Administration could maintain a structured register of every formal stage and deadline.

Even then, sanction frequency should not become a performance target. A healthy constitutional system does not seek a quota of board removals. The relevant measures are access, response, correction and confidence: could a serious concern enter; did the responsible institution explain its choice; did cross-community discussion occur; did the board alter an unlawful or damaging action; was a binding remedy available if it refused?

The two ALAC cases are valuable precisely because they are observable failures to advance, not proof of institutional failure overall. They locate friction at the second-support gate. Future cases may locate it elsewhere. Evidence should follow the stage, not a slogan about empowerment.

Ordinary accountability needs smaller remedies

The community powers were framed as safeguards of last resort. NTIA itself said before the transition that the hope was that such powers would rarely, if ever, need exercise. That is a defensible aspiration for recall and court action. It becomes dangerous if every concern is forced to choose between ordinary comment and a constitutional weapon.

Institutional accountability works better with a ladder of remedies. At the lower rungs, a Decisional Entity should be able to obtain a reasoned board response, request specified records, secure a time-bound implementation explanation or convene a public issue session. Mid-level remedies can include reconsideration, mediation, independent investigation and review. Rejection, removal, recall and litigation remain at the top.

Some of this ladder already exists. The bylaws provide inspection rights and allow three Decisional Entities jointly to trigger an independent investigation of a credible allegation of fraud or gross mismanagement of ICANN resources. Reconsideration and independent review are available under defined conditions. Public comment and periodic reviews provide additional scrutiny. The problem is not total absence but fragmentation and differing activation rules.

The 2025 dispute about ATRT4 illustrates the need for an intermediate remedy. ALAC wanted immediate reconsideration; ccNSO agreed the review cycle was troubled but preferred a pause to reassess it. A structured, time-bound joint examination of the disputed timing and remedy might have converted shared diagnosis into a narrower commitment. The all-or-nothing support gate recorded disagreement without requiring such a bridge.

Smaller remedies should not let one constituency obstruct operations indefinitely. They should require a concrete issue, public reasons and a response date. Their value would be to surface and resolve ordinary disagreement before it matures into the kind of extreme board conduct for which the Empowered Community is best equipped.

Transparency should expose decisions, not manufacture unanimity

Several practical improvements follow from the evidence without altering the constitutional bargain. Every Decisional Entity could publish one durable page setting out who may petition, the filing method, completeness requirements, internal decision-maker, voting rule, consultation route and expected timetable. The bylaws already require such procedures; common presentation would make them usable.

When a entity accepts, rejects or declines to support a petition, it should publish a short reasoned decision. The ccNSO did this in 2025, distinguishing concern about delays from disagreement with immediate reconsideration. The ASO and ccNSO supplied useful 2026 explanations. Applying the practice consistently would let observers tell substantive opposition from lack of time, standing or consultation response.

The Administration could publish a single event ledger for each action: board notice, petition window, accepted petitions, support deadline, entity positions, forum materials, final threshold and legal effect. The correspondence archive contains much of this material, but users should not need to reconstruct one procedure from scattered letters.

Entity reports should state participation boundaries. They may disclose how many eligible councils, members or regional bodies were invited and how many responded where lawful and meaningful, but should not imply that silence equals approval. They should preserve minority reasons where internal rules permit. Such records would support institutional learning without inventing a universal electorate.

Finally, complex actions could receive a support period calibrated to the practical time needed for five global institutions to consult. Any extension must preserve certainty for budgets and operations. A small increase or an automatic public orientation call during the support stage could improve informed coalition-building without reducing the final supermajority.

These reforms would not make the Empowered Community a member. They would make its designator model easier to see, enter and evaluate on its own terms.

The bargain is constraint without direct rule

The Empowered Community is neither a decorative advisory panel nor a parliament of Internet users. It is a legal association of five ICANN institutions with enumerated rights over a memberless corporation. Its powers are real: protected amendments need its consent, valid rejections nullify specified board actions, valid removal notices remove directors, and courts cannot dismiss it merely for lacking legal capacity.

Its democratic claim is more modest. Individuals and organisations act through Decisional Entities whose constituencies, internal rules and incentives differ. One entity can open several procedures. It generally cannot complete them alone. The 2025 and 2026 ALAC petitions show that a serious, public and properly framed challenge may still stop before collective deliberation because no second institution chooses to support it.

That friction was not a drafting accident. The mechanism was built to demand cross-community agreement before inflicting high-cost remedies. It protects ICANN from capture and sudden instability. It also protects board decisions from diffuse concern that cannot rapidly align institutional representatives. The same threshold is both safeguard and barrier.

The fairest judgment is therefore conditional. For extreme board conduct that is concrete enough to unite names, numbers, users, country-code managers and perhaps governments, the association supplies a credible escalation path and enforceable end point. For ordinary accountability - delayed reviews, contested priorities, weak explanations and implementation drift - its representation chains and coalition gates can be too coarse.

ICANN's post-transition legitimacy does not require pretending that the Empowered Community is a statutory member or a global electorate. It requires showing that those who can trigger its stages are identifiable, that institutions give reasons, that thresholds fit remedies and that lawful decisions bind the board. The community is empowered when that chain works. Where it stops, the exact stopping point should be public.

Sources