Summary

  • Internet Society’s bylaws permit its Board to appoint as many as three voting trustees by a two-thirds majority. The other twelve category seats belong to Organization Member elections, Chapter elections and IETF appointment.
  • A procedure adopted on 11 February 2026 limits the Board-appointed route to rare, extraordinary needs, requires the decision and rationale to be disclosed, and places recruitment with the Governance Committee and a professional search firm.
  • The same procedure says the search may or may not include a public call. It requires a common minimum candidate packet and extensive review, but the public record does not prescribe aggregate evidence showing how widely the declared gap was tested.
  • An exceptional-seat selection receipt can expose the search mode, criteria version, aggregate funnel, conflicts, recusals, vote denominator and result without exposing applicants, references, background checks or protected deliberation.

The smallest gate in the procedure

One conjunction controls who can enter the visible side of Internet Society’s exceptional trustee process. The Governance Committee and its professional search firm will choose the best recruitment method, the February 2026 procedure says, and that method “may or may not include a public call.”

That is not a drafting curiosity. A public call says that a qualified outsider can discover the opportunity and place a candidacy into the governed funnel. A targeted search says that the recruiter and committee define the discovery universe first. Both can produce an excellent trustee. Both can also be operated fairly. But they create different evidence about who could appear, who did appear and how confidently the chosen candidate answers the gap that justified bypassing the ordinary seat routes.

The procedure is unusually candid about the choice. It does not advertise an open process while quietly reserving the right to recruit by invitation. It expressly permits either route. The governance question therefore is not whether the text should be caught in contradiction. It is what public evidence should accompany a valid decision to keep the call closed.

A fourth route beside twelve category seats

Internet Society’s bylaws divide the ordinary Board structure among three selectors. Organization Members elect four trustees. Chapters elect four. The IETF appoints four through a process of its own choosing. Beside those twelve seats, the Board may appoint up to three more. It need not fill any of them, and a Board appointment requires a two-thirds majority of trustees then in office.

The special procedure explains why that fourth route exists. In rare cases, the Board may find a pressing skill or perspective missing from its composition. The published examples are concrete: nobody able to read and understand a balance sheet, every trustee coming from one country, or a significant decline in gender diversity. These are examples, not a closed list.

The sequence begins with a Board member asking that an appointment be considered. The Board identifies and discusses the need. A secret two-thirds vote is required to begin recruitment, and a trustee who wants to be considered must recuse. If the vote succeeds, the Governance Committee manages the search. The Board’s decision and rationale must be disclosed publicly.

That public rationale is important. It prevents an additional seat from appearing as an unexplained act of self-extension. It names the defect the new seat is supposed to repair. But a rationale is a hypothesis. “We lack financial literacy” does not prove that the candidate universe was designed to find the strongest financial oversight. “We lack a perspective” does not prove that the selector searched beyond its familiar network. The claim and the test are different institutional objects.

The procedure contains serious controls

It would be inaccurate to describe the Board-appointed route as an unstructured private choice. The procedure does substantial work.

The Governance Committee is to use a Board matrix covering perspectives, experience, skills and diversity. A professional search firm manages candidate identification and recruitment. Every candidate under consideration must provide at least a current CV, a statement addressing the selection criteria, references and a completed conflict-of-interest disclosure. The Committee and search firm screen applicants against established criteria, interview qualified candidates, conduct reference and background checks as appropriate, evaluate candidates against the Skills Matrix and assess potential conflicts.

The Committee then selects one finalist for each open seat. Its recommendation package must map the finalist’s qualifications to the Skills Matrix, explain how the candidate addresses the identified Board gap, summarize interviews and checks, and state alignment with Internet Society’s mission and values. The Board reviews the recommendation, discusses the candidate in Closed Executive Session and votes using the bylaw supermajority. A trustee under consideration must recuse from discussion and voting.

Those controls separate recruitment, evaluation and authorization. The search firm finds and tests candidates but does not appoint. The Governance Committee narrows and recommends but does not confer office. The Board decides, but only after it has publicly justified starting the exceptional route and has met a heightened voting rule.

The procedure also protects legitimate interests. Candidates may be unwilling to have an exploratory conversation become a searchable public fact. References and background checks contain personal information. A sitting executive may not be able to declare interest before knowing whether a serious process exists. Board deliberation about weaknesses and fit can require candour. Confidentiality is not an embarrassment to be eliminated; it is a design constraint.

Equal treatment is not an open entrance

The most tempting shortcut is to say that all candidates follow the same thorough process, so the search is public enough. That confuses treatment inside the funnel with access to the funnel.

A common application packet is evidence of procedural equality among people the search has reached. It does not show how those people were found. A Skills Matrix can discipline evaluation. It does not identify which regions, professions, networks or institutions were searched. A professional recruiter can expand the field far beyond the Board’s contacts. It can also run a precise, invitation-only search whose breadth is visible only to the client.

None of these observations proves bias. They reveal a missing join. The public record tells an outsider why the Board says an exceptional seat is necessary. It may later identify the successful appointee. Between those points sits the evidence that tests whether the route matched the reason.

If the declared gap is balance-sheet literacy, the relevant public question is not who the losing candidates were. It is whether the criteria gave financial oversight real weight, whether the search was open or targeted, how many candidates reached eligibility and interview stages, and whether the finalist’s disclosed qualifications map to the declared need. If the gap is geographic concentration, aggregate coverage by region may matter. If it is gender diversity, the measure should be privacy-safe and suited to the actual claim. One universal diversity dashboard would be worse than a receipt tied to the stated gap.

One finalist is a recommendation state, not a funnel count

The words “one finalist per open seat” can be read in two opposite ways. They can mark responsible committee work: the Board receives a developed recommendation rather than rerunning interviews in plenary. Or they can look like a single-candidate ratification if no evidence survives about what preceded the recommendation.

The procedure itself supports the first reading. It refers to all candidates, screening, interviews, checks and evaluation. One finalist therefore does not mean only one candidate existed. But the distinction should not depend on a sympathetic reader reconstructing verbs from several paragraphs. A public aggregate can make it durable: candidates identified, candidates who entered formal consideration, eligible candidates, interviews completed, candidates advanced, and one finalist recommended.

The counts need definitions. A name produced by a recruiter is not an applicant. An exploratory contact is not consent to candidacy. A person who submits the minimum packet is not necessarily eligible. An interview does not imply endorsement. Without states, a large top-line number can manufacture an appearance of breadth just as easily as silence can manufacture suspicion.

The ordinary election is a comparator, not a template

Internet Society’s 2026 ordinary cycle made another kind of evidence visible. It published the nomination window, eligibility cutoff, initial and final slate dates, candidate forum period, voting dates, challenge period and finality date. The nominee page listed five candidates for two Chapter seats and three candidates for one Organization Member seat. The final-results notice named the selected Chapter, Organization Member and IETF trustees and described the challenge path.

That process has a different principal and purpose. Chapter voters and Organization Member voters choose their allocated trustees; the IETF uses its own appointment route. A Board-appointed seat exists precisely because those ordinary paths may leave a specific composition gap. Copying an election mechanically could weaken the exceptional tool. A public popularity contest is a poor way to recruit a rare audit skill, and a fully public longlist may deter qualified candidates.

Yet the ordinary cycle demonstrates a useful principle: a selection route can publish its states without publishing every internal judgment. Dates, eligibility, slate, forums, voting and challenges create a chain that readers can audit. The exceptional route needs a different chain, but not no chain.

A receipt that preserves the private room

The right public object is smaller than a dossier and stronger than a press release. It is an exceptional-seat selection receipt.

Before recruitment, the receipt would identify the bylaw authority, number of seats considered, declared extraordinary gap, Board vote to proceed, any recusal count, and the published rationale. It would identify the Skills Matrix and selection-criteria versions used for this search.

At the search-mode decision, it would state public call, targeted search, or hybrid, name who approved the mode and give a bounded reason. A targeted search might be justified by candidate confidentiality, a scarce skill, time, market sensitivity or the need to approach people not actively seeking a role. The reason is evidence of the trade-off, not a presumption of wrongdoing.

During recruitment, only aggregate states need be public: identified, contacted, consented to consideration, complete packet, eligible, interviewed and advanced. Counts smaller than a safe privacy threshold could be suppressed or banded. The receipt could record whether references, background checks and conflicts were completed without publishing their contents. It could state the number of recusals without naming a private conflict unless disclosure is independently required.

At authorization, the record would state that one finalist was recommended, link a public qualifications-to-gap summary, identify the eligible vote denominator and required affirmative count, and publish the outcome without exposing individual secret ballots. It would record the term, effective date, whether the person is a first-term or reappointment candidate, and any correction or superseding decision.

This design does not invite the public to veto the candidate. It does not give the search firm a mandate. It does not turn applicants into representatives of every person who shares their geography or profession. It simply preserves the proof that the exceptional route remained tied to the exception that activated it.

The denominator deserves its own field

The voting text shows why a receipt must record mechanics, not merely say “two thirds.” The procedure’s opening vote requires two thirds of trustees then in office, excepting trustees recused from voting. The final appointment vote refers back to the bylaws’ two-thirds majority of trustees then in office, while separately requiring a trustee candidate to recuse from discussion and voting.

The public procedure does not explain in the same sentence how a recusal affects the final denominator. This article does not choose an unpublished interpretation. It asks the Board to state the operative denominator and required affirmative count before the ballot, then preserve both in the result. A secret ballot can remain secret while its authorization arithmetic is public.

That distinction is not clerical. With a close vote, changing the denominator by one can change the threshold. A resolution that says only “approved by two thirds” forces later readers to guess which population was counted. A governance record should never require arithmetic folklore.

What the evidence does not show

No source reviewed for this article says that Internet Society is now recruiting a Board-appointed trustee. None proves that the no-public-call option has ever been used under the 2026 procedure. The public source set does not contain the Skills Matrix itself, but that does not mean the matrix is absent internally. The current Board headcount does not establish how any particular trustee reached office, and no such inference is made here.

Nor does the procedure appear, on its face, to transfer appointment power to the recruiter or Governance Committee. The Board retains the final vote. The February resolution was adopted to create a clearer process around an existing bylaw power. The proper critique therefore is not that no rules exist. It is that a strong internal chain can still produce a thin public proof surface.

Heng Lu’s challenge to multistakeholder legitimacy is useful at exactly this point. A selection method may produce skill and sound judgment without producing a mandate from everyone affected by the institution. Conversely, the absence of a public call does not make an appointee illegitimate if the constitutive authority validly provides another route. The disciplined question is narrower: who selected, under which rule, to solve which declared problem, with what evidence, and with which liability and correction path?

Keep the exception; make it legible

Internet Society’s Board-appointed route answers a real governance need. Constituent selectors cannot guarantee every annual mix of skills and perspectives. Vacancies and changing risks do not arrive on the election timetable. A Board sometimes needs a bounded way to repair its own capacity.

The February procedure improves that power. It makes the seat optional, describes exceptional triggers, requires a public rationale, separates recruiter, evaluator and decision-maker, imposes common candidate materials, uses conflicts and recusals, limits the recommendation and demands a supermajority.

Its weak point is not secrecy in general. It is the unjoined middle between public reason and public result. The sentence allowing no public call should remain if the Board believes targeted recruitment is sometimes necessary. But every time that option is chosen, the choice should create a receipt.

A public call is one form of evidence. It is not the only form. A non-public search therefore needs another: a durable, aggregate account proving that the exceptional seat was searched, tested and authorized for the exception named at the start.

Sources

  1. Lu Heng, “The Multi-Stakeholder Mirage—How the Multi-Stakeholder Model Turned Attendance Into Mandate”
  2. Lu Heng, “On Who Gets to Speak for a Continent, a Community, or ‘The End User’”
  3. Internet Society, Amended and Restated By-Laws
  4. Internet Society, Selection Procedures for Board-Appointed Trustees
  5. Internet Society, Resolution 2026-1
  6. Internet Society, Minutes of Board Meeting 192
  7. Internet Society, Governance Committee
  8. Internet Society, 2026 Board Elections
  9. Internet Society, 2026 Nominees
  10. Internet Society, Final 2026 Election and IETF Selection Results
  11. Internet Society, Procedures for Selecting Trustees
  12. Internet Society, Transparency