This series does not repeat old work. It asks a narrower and more consequential governance and political-institutional question: once an ICANN-supported regional process began touching the governance of an independent Regional Internet Registry, what duty did ICANN have to warn, correct, and record the boundary before the process acquired institutional momentum?

The issue is no longer only whether ICANN was involved. The issue is whether ICANN’s involvement was bounded early enough, clearly enough, and publicly enough to prevent its institutional presence from becoming implied legitimacy for a process affecting AFRINIC governance.

That question matters beyond Africa. If ICANN can support a regional Internet governance process in one region and later disclaim responsibility when that process drifts into RIR governance, the same ambiguity can travel anywhere. AFRINIC is the case study. The lesson belongs to the system.

Why the ICANN—Smart Africa—AFRINIC episode is not about authorship alone?

The most important question in the ICANN—Smart Africa—CAIGA case is not whether ICANN wrote CAIGA, that question has already consumed enough oxygen. Nor is the central question whether Smart Africa has the right to organize Internet governance discussions across Africa, it does. Nor even whether governments may participate in Internet governance, they may, they are, and in many cases they should.

The harder question is different:

When ICANN funds, facilitates, attends, or lends institutional credibility to a regional Internet governance process, what duty does ICANN have when that process begins drifting toward the governance of an independent Regional Internet Registry?

That is the question still insufficiently answered.

The ICANN—Smart Africa—AFRINIC episode is not merely an African controversy. It is a global warning. It shows how an ICANN-supported regional initiative can begin as capacity building, policy dialogue, or continental coordination, then gradually acquire the shape of institutional redesign. By the time the affected community sees the full architecture, the proposal may already have passed through meetings, presentations, working groups, political language, funding structures, and institutional associations.

At that point, ICANN can say it did not author the proposal, it can righteously say it did not endorse it, it can claim that it did not intend the outcome, but those answers do not fully resolve the governance problem.

In Internet governance, legitimacy is not created only at the moment a proposal is formally submitted, it begins at an earlier point: when the room is funded, the agenda is framed, the participants are selected, and when institutional logos appear, then when the technical vocabulary is supplied, and when silence allows ambiguity to mature.

This is why the issue is more serious than solely facilitation without correction and extends further to facilitation without responsibility.

By that, I mean the situation where an institution funds, supports, attends, or gives credibility to a process, but later disclaims responsibility for the governance effects of that process by saying it did not author, control, request, or endorse the final output. In ordinary grant-making, that defense may be sufficient. In Internet governance, especially where ICANN is involved and where the output touches the governance of a Regional Internet Registry, it is not enough.

ICANN is not a passive foundation nor just another sponsor of a regional workshop. It is the Global Coordination Institution associated with the unique identifier system and with the recognition framework under which Regional Internet Registries exist. When ICANN participates in a process that later touches RIR governance, its responsibility is not limited to whether it drafted the offending paragraph. It must also answer whether it maintained the boundary that its own role made necessary, and that boundary is ICP-2.

The Authorship Red Herring

In governance systems, responsibility does not arise only from textual authorship. It may also arise from facilitation, funding, institutional association, delayed correction, and failure to warn. Authorship asks who wrote the text, while “Boundary duty” asks who stood in the room when the text crossed into institutional territory, and what they did when that crossing became visible.

This is why ICANN’s answer, while technically important, remains institutionally incomplete. It denies authorship among other things, but it does not fully answer the duty of “contemporaneous warning”. If ICANN was present to provide ICP-2 reference points, then the central question is whether it merely pointed to ICP-2, or whether it used ICP-2 as a guardrail when the Smart Africa process entered AFRINIC-governance territory. This distinction is not semantic, but constitutional.

A global coordination institution does not need to draft a proposal in order to lend it force. Sometimes it is enough to fund the room, sit at the table, provide technical vocabulary, and remain silent while the room begins to treat political architecture as governance reform.

That is why the authorship question, standing alone, is a red herring. The deeper question is not whether ICANN wrote or endorsed AIGA. The deeper question is whether ICANN allowed its institutional presence to make CAIGA appear more procedurally legitimate than the AFRINIC community had yet made it.

The Rule of Contemporaneous Warning

Neutrality is not a post-crisis disclaimer.

It is not enough for a coordinating institution to say, after controversy has matured, that it did not intend, request, pre-evaluate, author, or endorse the disputed output. Those answers may clarify legal distance, but they do not repair institutional drift once legitimacy has already been transferred through funding, presence, facilitation, and silence.

In any governance system built on delegated trust, the warning must be contemporaneous. It must appear when the boundary is crossed and not when the political cost of correction has already increased. If ICANN was present in a process to provide reference to ICP-2, then its responsibility was not satisfied by quietly pointing to public documentation. The relevant question is whether ICANN issued a clear warning when the Smart Africa process began moving from general Internet governance capacity building into AFRINIC-specific governance redesign.

That warning didn’t need to be hostile, nor to accuse Smart Africa, nor needed to halt regional dialogue. The whole issue was reducible to mark the institutional boundary: RIR governance cannot be re-authored through governmental or intergovernmental coordination without the affected RIR community’s prior, transparent, and meaningful participation. That is the line ICANN had to draw.

The duty to warn is heightened where the affected institution is fragile. AFRINIC was not operating under normal conditions. It was under receivership, without ordinary board continuity, with election legitimacy under contestation, and with external actors already framing stabilization as reform. In such a setting, institutional silence cannot be considered weightless, rather it becomes part of the political economy of the room.

The deeper failure, therefore, is not only that ICANN later had to clarify its non-endorsement. The deeper failure is that the clarification came after the architecture had already travelled, and this failure continues while the same proposal is not being validated.

In Internet governance, timing is substance. A warning issued before institutional drift can preserve legitimacy while A disclaimer issued after drift has matured merely records distance.

ICANN Was Not Distant From AFRINIC’s Crisis

ICANN’s own record confirms that it was not distant from the AFRINIC crisis. In March 2025, ICANN wrote to AFRINIC’s appointed receiver, Mr. Gowtamsingh Dabee, explaining AFRINIC’s role in the numbering system. ICANN stated that, prior to the removal of the earlier receiver, it had offered “neutral and independent assistance” to him at his request, and that ICANN’s appointed counsel in Mauritius had already held an initial meeting with the new receiver. The letter also stated that it could be shared with the courts in Mauritius.¹

That matters because ICANN knew AFRINIC was not operating in normal institutional weather. AFRINIC was under receivership, with neither a board nor a CEO. Its governance legitimacy was contested and its role as the RIR for Africa and parts of the Indian Ocean was not simply corporate; it was infrastructural. Any external process touching AFRINIC governance therefore required a higher standard of boundary discipline, not a lower one.

The NRO also had already recognized, as far back as March 2023, that AFRINIC’s lack of a functioning board required elections to restore quorum, and that a free and fair AFRINIC election required effective activation of the AFRINIC membership and Community toward informed participation. The same letter referred to the need for community-based solutions and to reducing opportunities for capture by any single interest group.²

That standard is important. The institutional standard was not cured by “hold an election.” It was informed participation, community activation, and protection against capture. Those principles are broader than ballot mechanics.

Yet by 2025, the Smart Africa process was no longer confined to general Internet governance dialogue. At AIS 2025 in Accra, the relevant session was titled “High-Level Multistakeholder Forum on AFRINIC Reforms and the Continental CAIGA Framework.” The agenda listed high-level interventions from ICANN, ISOC, etc.., followed by “AFRINIC Reform Process, Status Update and Next Steps” by Smart Africa Ad Hoc Committee support, and later a “Presentation of the CAIGA Framework.”³

This was not merely a generic Internet governance capacity-building workshop. The words were on the page: AFRINIC reforms and CAIGA were being discussed in the same institutional setting.

That is where the boundary question begins.

ICANN Answered “Authorship”, Not “Warning”

ICANN later clarified that it had entered into a Memorandum of Understanding and a Project Agreement with Smart Africa, and that it committed US$40,000 for Smart Africa to develop the Internet Governance Blueprint. ICANN further stated that it did not enter the MoU or Project Agreement with the intention that the Blueprint would propose a new governance model for the RIR serving Africa and the Indian Ocean; that it did not request Smart Africa to deliver proposals on AFRINIC governance; and that it did not pre-evaluate the CAIGA proposal.[4]

Those statements may answer the authorship question. They do not answer the warning question.

Once CAIGA appeared inside the Blueprint, intent stopped being the central issue. The central issue became correction. If a project funded and supported by ICANN produces a proposal touching the governance of an RIR, the question is not only whether ICANN intended that outcome. The question is what ICANN did once the outcome appeared.

Did ICANN warn Smart Africa that RIR governance could not be politically redesigned outside the affected RIR community?

Did ICANN place on record that any proposal affecting AFRINIC’s governance must be transparently authored by, and supported by, the AFRINIC community under ICP-2?

Did ICANN tell the community, before public controversy erupted, that the Blueprint had produced an RIR-governance proposal whose compatibility with ICP-2 was doubtful?

Did ICANN require Smart Africa to correct any public statement suggesting broader ICANN involvement in CAIGA than ICANN believed was accurate?

If such warnings exist, ICANN should publish them.

If they do not exist, then the problem is not communicative inconvenience, it is institutional boundary failure.

ICANN’s later response says that, as currently set out in the Blueprint, it was “not clear” to ICANN that CAIGA met the ICP-2 test, and that if any proposal to modify AFRINIC’s governance structure were to come to ICANN, ICANN would neutrally and objectively evaluate it. ICANN also stated that CAIGA would not receive priority or exemption from ICP-2 principles simply because it emerged from a project that used ICANN funding.[5]

But this answer exposes the deeper problem.

Why was ICANN waiting for a proposal to “come to ICANN” when the proposal was already moving through ICANN-associated spaces, Smart Africa processes, public presentations, and an ICANN-funded project?

ICANN positioned itself as a future evaluator of CAIGA while already being part of the ecosystem through which CAIGA gained visibility. That is not neutrality in any meaningful governance sense. That is delayed accountability.

A proposal does not become powerful only when it arrives formally at ICANN’s door. It becomes powerful when it is presented in a room with recognized institutions, when it is framed as regional consensus, when governmental actors are involved, when technical institutions appear nearby, and when affected communities will be asked to respond after the architecture has already been drafted.

That is why publishing the ICANN—Smart Africa partnership was not enough. A public partnership announcement is not the same as public disclosure of the governance consequences of that partnership.

The issue is not whether the MoU existed on a website. The issue is whether the AFRINIC community was given timely, meaningful visibility into the Blueprint’s RIR-governance implications before those implications gained institutional momentum.

ICANN’s response further states that its coordination with Smart Africa was focused broadly on Internet governance in Africa, not CAIGA; that Smart Africa and its consultant included CAIGA as part of the broader Blueprint; and that when ICANN was approached to provide substantive inputs on CAIGA, it declined to do so, referring instead to publicly available documentation, including ICP-2.[6]

Again, that answers only part of the question.

Referring privately or internally to ICP-2 is not the same as warning publicly that a proposal is entering RIR-governance territory. Refusing to draft CAIGA is not the same as correcting the environment that allowed CAIGA to acquire credibility through an ICANN-funded process, and presentations on the sides of ICANN meetings.

A boundary is not maintained by declining to write the proposal. A boundary is maintained by saying clearly, while the proposal is still moving, that the proposal is approaching or crossing a constitutional line.

This is the duty ICANN has not yet fully explained.

It is true and valid that It was not ICANN’s job to: govern AFRINIC, run AFRINIC elections, nor to police every Smart Africa discussion. But once ICANN funded, participated in, or lent institutional presence to a process that produced governance proposals affecting an RIR, it became ICANN’s job to ensure that its own presence was not converted into implied legitimacy.

ICANN was not required to govern AFRINIC, it was rather required not to let its own institutional association blur who may govern AFRINIC. That is the difference, and that is the missing test.

Soon Part 2 | When Non-Endorsement Is Not Enough: CAIGA, ICP-2, and the Risk of Legitimacy Transfer.

Background Reading by the Author

For earlier analysis of the ICANN—Smart Africa cooperation record, CAIGA, Smart Africa’s public reform positioning, and the broader legitimacy question around AFRINIC, see the author’s prior writings on Medium and CircleID listed below.

1. Inquiry About the “Joint Work”, ICANN—Smart Africahttps://medium.com/p/99a74903804c
2. When Neutrality Isn’t Neutral: ICANN, Smart Africa, the “BluePrint” and “CAIGA”https://circleid.com/posts/when-neutrality-isnt-neutral-icann-smart-africa-the-blueprint-and-caiga
3. Consensus or Capture? Smart Africa’s September 13, 2025 Communiquéhttps://medium.com/p/499e5bffdd45
4. The Internet Legitimacy Gap: When Governance Outgrew Its Architecturehttps://circleid.com/posts/the-internet-legitimacy-gap-when-governance-outgrew-its-architecture
5. Who Authorizes Legitimacy if the Community Is Absent in the Redefined Procedural Spacehttps://circleid.com/posts/who-authorizes-legitimacy-if-the-community-is-absent-in-the-redefined-procedural-space