Summary

  • Official IGF 2025 participation statistics report 3,435 onsite entities, more than 6,000 online entities and total participation above 9,435. Online entities therefore made up at least about 63.6 per cent of the stated combined total.
  • Remote numerical majority widened access but did not create a voting majority. The IGF has no entity ballot for themes, workshop selection, annual messages or external policy, and most online attendees entered after the agenda had been set.
  • The MAG and Secretariat screened 400 workshop proposals and selected about 51 for the programme. The host country supplied facilities, a local organizing structure and high-level convening resources. Session organizers and moderators controlled the conversion of proposals, speaker lists, chat and raised hands into visible programme time.
  • The official record reports 33 remote hubs in 21 countries, heavily concentrated in Africa, and estimates that about 1,700 people benefited from the hubs. Hubs can add local context and collective presence, but the published totals do not reconcile hub entities with the 6,000-plus online figure.
  • Hybrid legitimacy should be assessed through agenda access, speaking time, moderation outcomes, language and bandwidth, time zones, organizing roles, session acceptance, informal access and downstream response, not by treating online logins as delegated authority.

The majority that was not a majority rule

More than 6,000 online entities against 3,435 onsite entities creates an arresting image. The digital room was larger than the physical one. At minimum, the online share was about 63.6 per cent of the combined total, and the remote-to-onsite ratio was about 1.75 to one. For a forum about the Internet, this looked like a natural experiment: what happens when the people beyond the venue become the numerical majority?

The first answer is encouraging. The meeting did not require global interest to fit inside NOVA Spektrum. Participation could extend to people who lacked travel budgets, faced visa barriers, could not leave care work, had a disability that made the journey difficult or simply wanted to join one relevant session rather than spend a week at a conference. Thirty-three remote hubs added collective settings in 21 countries. The virtual environment included entry points for sessions and online versions of exhibition spaces.

The second answer is institutional. Numerical majority has political meaning only when rules attach decision rights to numbers. The IGF did not ask all entities to vote on themes, workshops, messages, appointments or policy. One online entity did not receive one unit of agenda power, just as one badge holder did not. The Forum remained a dialogue institution whose programme was assembled through calls, advisory review, Secretariat administration, host-country support and organizer decisions.

Remote majority therefore describes access mode. It does not describe the distribution of constitutional authority. The distinction should have been obvious, yet conference language often slides from "most entities were online" to "the online community shaped the event" without identifying a mechanism between the two claims.

Lillestrøm should be judged as a major access achievement and a test of whether access reached deeper layers. Did remote entities propose themes? Were their workshop proposals selected? Did they organize, speak and moderate? Were their questions heard? Did their concerns survive the reports? Could they form relationships that reached institutions with power? The official headcount begins that inquiry. It does not complete it.

The official denominators do not reconcile cleanly

Before analysing influence, the participation numbers need a careful reading. The official web page says the meeting received 8,833 registrations. It then reports 3,435 onsite entities and more than 6,000 online entities, producing a total above 9,435. The final IGF 2025 Summary Report describes 9,435-plus stakeholders registered from 85 per cent of UN member states, alongside the same onsite and online figures.

Those statements cannot all use "registrations" in the same way. A combined participation total above 9,435 is larger than 8,833 registrations. The discrepancy may reflect multiple registration systems, late additions, remote-platform identities, group participation through hubs, an editorial label or a revised total. The published materials do not provide a reconciliation.

Other composition figures also vary. The web page's narrative gives government 35 per cent and civil society 34 per cent, followed by private sector at 14 per cent and technical community at 10 per cent. An accessible label attached to the page's graphic gives civil society 32 per cent, government 27 per cent, private sector 19 per cent and technical community 15 per cent. The final report says civil society was the largest group at 32 per cent. Regional figures differ between the web narrative and its graphic label as well.

These are not reasons to discard the record. The Summary Report and participation page agree on the central mode totals: 3,435 onsite and more than 6,000 online. They are reasons to use thresholds and cite the specific presentation. They also demonstrate why a hybrid meeting needs a denominator note that defines people, registrations, platform accounts, hub entities and stream viewers.

The official report separately calls the 6,000-plus group "Zoom delegates" who joined sessions remotely. Livestream services recorded more than 22,000 views. A view is not a delegate; a delegate may join multiple sessions; a person may use both onsite and online facilities. Without published deduplication and activity rules, the precise remote share should not be stated beyond the official threshold.

The safest conclusion remains strong: by the Forum's reported method, online participation substantially exceeded onsite participation. The unsafe conclusion is that a measured electorate of 6,000 distinct remote people approved any programme or recommendation.

Access moved; agenda formation had already happened

The annual meeting began on 23 June 2025. Its agenda began months earlier. A public call for thematic inputs and stocktaking comments informed the first consultations. The MAG then agreed the overarching architecture and four subthemes: digital trust and resilience, sustainable and responsible innovation, universal access and digital rights, and digital cooperation. Session proposers submitted applications by March. MAG members evaluated workshops, and the selection was approved in April.

This sequence is not secret or inherently exclusionary. The call for session proposals was open, and any person or group could submit a workshop. The thematic-input process was also open. Proposal criteria rewarded relevance, content, policy questions, engagement, hybrid design and diversity. Open consultations could be attended online.

But access to a June session is not the same as access to the January-to-April agenda. Most of the reported 6,000-plus online entities were counted at the meeting. The statistics do not say how many took part in thematic calls, consultation meetings or proposal writing. A person can be admitted to every June room and still encounter only issues selected by an earlier and smaller population.

This temporal inequality matters because agenda power is path dependent. Once four subthemes are chosen, proposals must fit them. Once a workshop deadline closes, later entities can ask questions but cannot claim a scarce session. Once speakers are confirmed, an online attendee competes for audience time rather than framing the opening evidence.

The remote majority did not arrive too late by accident. Annual programmes need advance preparation. The governance question is whether the institution measures participation across the whole year instead of using annual-meeting totals to describe the inclusiveness of earlier choices.

A transparent report would show how many thematic inputs, proposals and evaluation-stage observers came from people who later participated remotely. It would show whether remote-first and low-resource organizers advanced from input to accepted session. Without that chain, a large online audience demonstrates access to a completed agenda more clearly than access to agenda formation.

Fifty-one selected workshops reveal the scarcity

The IGF 2025 workshop page says the MAG selected 51 workshops from 400 proposals. The participation page later lists 52 workshops among 262 sessions, a one-session difference that may reflect a programme change or category count. Either way, only about one proposal in eight obtained a workshop slot.

Scarcity turns evaluation into power. The 2025 rules assigned weight to topic, content, policy questions, engagement and hybrid strategy, and diversity. MAG members were grouped to review proposals, individual scores were averaged, and the Secretariat prepared the results for group evaluation and final approval. The process sought balance across themes rather than applying a mechanical rank alone.

These are defensible programme functions. A meeting cannot place 400 workshops into 51 rooms without selection. Quality, duplication, diversity and schedule all require judgment. The important point is that the judgment belonged to the MAG and supporting Secretariat process, not to the eventual entity majority.

Hybrid quality entered the selection criteria. That was valuable: organizers had to explain how online and onsite entities would engage on equal terms. Yet a criterion is a promise assessed from an application. It does not show whether remote entities later received equal floor time, whether audio worked, whether questions were voiced or whether online speakers were treated as central rather than supplementary.

The evaluation population also differed from the attendance population. MAG members are appointed by the UN Secretary-General and serve in personal capacity while maintaining stakeholder links. They do not receive instructions from six thousand remote attendees. The MAG terms of reference give them responsibility for themes, programme, schedule, main sessions and workshop selection.

Remote majority did not dissolve this advisory structure. It made the structure more important because a larger and more geographically dispersed public depended on a small group to allocate common attention.

The MAG controls programme power, not the policies discussed

Agenda power should be stated accurately. The MAG can decide which proposed workshop enters the official programme, where sessions sit, which broad themes receive prominence and how main sessions are organized. These decisions affect visibility, networking opportunity and the likelihood that an idea appears in annual synthesis.

The MAG does not thereby gain substantive jurisdiction over artificial intelligence, connectivity, cybersecurity, rights, standards or platform governance. An accepted workshop is not a policy approval. A rejected proposal is not a judgment that the underlying issue is invalid. The responsible government, company, operator, standards body or other institution retains decision authority.

This bounded power still needs accountability. A remote-heavy forum should publish proposal acceptance by organizer mode, region, stakeholder group, gender, newness and development context where data quality permits. It should show how many accepted sessions included genuinely remote organizers and speakers rather than adding an online moderator to an onsite design.

It should also expose concentration. If the same organizations repeatedly submit successful proposals, serve on panels and participate in consultations, the annual crowd can grow while agenda authorship stays narrow. Repeat expertise is valuable, but a pattern should be visible and justified.

The online majority could influence future MAG choices through stocktaking, consultations and new proposals. That is prospective influence, not retrospective control of the Lillestrøm programme. The distinction matters when the Forum describes its output as community-driven. "Community" should refer to a documented process and active contributors, not automatically to every person counted in June.

Agenda legitimacy comes from open calls, reasoned selection, plural evidence, conflict safeguards and correction. It does not come from attaching the whole attendance number to choices made by an appointed advisory group.

The host controlled enabling resources, not every session conclusion

Norway's host role gave the meeting physical reality. The IGF's hosting requirements expect a government to provide a venue capable of hybrid participation, local transport and accommodation conditions, political support, a local multistakeholder organizing committee and a high-level host-country co-chair. Host governments cover most annual organizational and conference costs under an agreement with the United Nations.

That is substantial power over enabling resources. Venue capacity determines how many rooms and entities can be accommodated. Technical facilities affect whether remote speakers can be heard. Security, accreditation, transport and local information affect access. A host can convene senior officials and draw attention to high-level moments. Ceremony and political visibility are not distributed by an online entity count.

It would be wrong to infer that Norway selected every workshop or dictated session conclusions. The formal selection materials assign workshop evaluation to the MAG and screening to the Secretariat. Funding and facilities create potential influence; they do not prove content control. The relevant safeguard is a visible division of responsibility.

Host influence also appears through location. The official statistics put Norway first among both onsite and online countries. Western Europe and Others was the dominant reported regional group, and more than 70 per cent of entities came from countries classified as developed. A host in an affluent, highly connected region can produce excellent technical conditions while still shaping whose institutions find participation easiest.

The meeting provided travel support to 130 people from developing countries and grants to remote hubs. Those interventions matter because nominally open access does not equal affordable access. They also show that host and Secretariat resource decisions determine which inequalities are partially corrected.

Remote majority did not eliminate host power. It made host investment in platforms, room integration and support a condition of participation for most of the counted public.

Moderators were the bridge and the gate

The official participation guidance gave online moderators a clear role: welcome remote speakers and attendees, monitor the virtual meeting, convey chat questions to the onsite moderator and respond to queries. This role was necessary. A hybrid room contains two attention systems, and someone must connect them.

The bridge is also a gate. A remote entity can raise a virtual hand or type a question, but the intervention reaches the physical floor through moderator attention, available time, technical quality and the onsite moderator's choice. The entity may not know whether a question was unseen, deprioritized, duplicated or excluded for time.

The 2025 stocktaking record contains both praise and warning. Contributors described reduced travel and visa barriers, online questions, captioning and live access. They also reported sessions in which remote interventions were delayed or deprioritized, inconsistent moderation, connectivity failures, weak multilingual support, time-zone problems and limited online networking. These are submitted experiences rather than an independent representative survey, but they identify the mechanisms a formal evaluation should measure.

Equality cannot be inferred from the existence of an online moderator. Reports should state how many remote hands were raised, recognized, voiced and recorded; how waiting time compared with the room; whether online speakers lost time to connection problems; and whether sessions had a dedicated moderator throughout. Aggregate reporting can protect individual privacy.

Design can reduce gatekeeping. A session may alternate onsite and online questions, reserve equal response windows, let remote entities speak directly when possible, publish unanswered questions and include remote contributors in the final summary. It can schedule online-first segments or use small mixed groups. These are choices about power, not cosmetic technical features.

Lillestrøm's numerical majority depended on human intermediaries to become audible. The institution should judge hybrid success at that conversion point.

Time zones turned simultaneous access into unequal convenience

Lillestrøm operated on Central European summer time. That placed live daytime sessions in workable hours for much of Europe and Africa, early or very early for parts of the Americas, and late afternoon or evening across much of Asia. No single location can be convenient globally.

The problem is cumulative. A entity joining one session may accept an inconvenient hour. An organizer, speaker or person seeking broad influence must follow multiple days, attend consultations, respond to chat and join informal calls. Time-zone burden compounds with care work and employment. People paid to participate can adapt more easily than volunteers or community members.

Recordings and transcripts improve knowledge access after the event. They do not provide a live floor, the ability to answer a misstatement or entry into a relationship forming in real time. Asynchronous comments can help only if organizers commit to reading, answering and preserving them.

A remote-majority forum should publish schedule impact by time zone. It could rotate recurring main sessions, repeat selected discussions, collect questions before the event and hold post-session responses at different hours. It should distinguish synchronous participation from later viewing when reporting engagement.

Time-zone equality is impossible in one global week. Procedural fairness is possible if burdens rotate and asynchronous participation has a documented route into reports and follow-up. Lillestrøm expanded availability, but the total does not show how many remote entities could use the live opportunities at reasonable hours.

Bandwidth and language shaped the remote public

Online access removes distance only for people with a device, electricity, sufficient connectivity and a usable language. The IGF platform provided video meetings, chat, transcription and streaming. Remote hubs could pool connectivity and local facilitation. These features widened access considerably.

They also selected for connected populations. A person facing unstable power or expensive mobile data cannot be treated as equally present because a link exists. Video-heavy environments impose higher costs than text or audio alternatives. A virtual three-dimensional venue may improve navigation for some and burden older devices or low-bandwidth connections for others.

Language creates another hierarchy. Main institutional materials may be multilingual, but many workshops operate in English without full interpretation. Automated captioning helps and can mistranscribe accents, names and technical terms. A entity who understands enough to listen may not be able to formulate a fast intervention in the working language before the moderator moves on.

These constraints affect agenda influence more than audience reach. A person can count as online after entering a session while remaining unable to speak. The statistics should therefore pair connection with participation outcomes and publish the languages in which sessions, captions and interpretation were available.

Remote hubs are a promising response because a local group can share infrastructure, discuss in a familiar setting and formulate collective interventions. They can also reproduce a local gate if one institution chooses attendance, controls equipment or speaks for everyone. Hub governance belongs in the evaluation.

The Internet made remote majority technically possible. Unequal Internet conditions shaped who could use it politically.

Thirty-three hubs created collective access and a counting puzzle

The final report says 33 remote hubs in 21 countries supported an estimated 1,700 stakeholders. It reports 76 per cent of hubs in Africa, 18 per cent in Asia Pacific, 3 per cent in Latin America and the Caribbean and 3 per cent in Western Europe and Others, while the participation web page gives Africa as 67 per cent. Seven hubs received grants.

The concentration in Africa is significant. Hubs can make a global meeting locally social. Entities discuss the relevance of a session, share connectivity and form relationships that persist after the screen closes. They can bring national or regional context into a global exchange more effectively than isolated logins.

The numbers are not fully reconciled. If 1,700 people benefited from hubs, were all included among the 6,000-plus Zoom delegates? Did hub entities register individually? Were some counted as a group? Does "benefited" include viewers who did not enter a meeting room? The published report does not say.

This ambiguity matters because hubs may account for a substantial part of the remote total. It also affects geographic analysis. A connection may appear under a hub account or local facilitator rather than each entity's country. Conversely, individually registered hub members may be counted both in platform and hub descriptions if the totals are not explicitly separated.

The answer is not to stop hubs or withhold estimates. It is to publish a hub ledger: location, organizer type, open recruitment method, entities, individual registration rule, sessions followed, interventions submitted, languages used, support received and follow-up. Aggregate and consent-based reporting can protect people at risk.

Hubs show how remote participation can become more than solitary viewing. They also show why a platform count alone cannot describe the social form of access.

Remote attendance did not reproduce the corridor

Formal sessions are only one layer of an annual meeting. Onsite entities encounter one another in queues, breaks, meals, exhibition areas, bilateral meetings and evening events. They can make an introduction in minutes, repair a misunderstanding after a panel or learn that an institution is about to open a consultation.

Remote entities may have chat, virtual booths and scheduled bilateral facilities. These can be valuable, especially for people who would otherwise have no contact. They do not automatically reproduce ambient discovery. An online entity generally needs a link, invitation or deliberate matching process. Leaving a session often means leaving the social space.

Research beyond the IGF supports this distinction. Duncan and Shean's study of online conference communication notes that virtual access can broaden participation while reduced informal contact limits knowledge transfer and network development. A 2024 mixed-methods study of a virtual food-systems conference found lower travel and visa barriers alongside weaker interpersonal connections and continuing inequality in Internet access. These studies are not measurements of Lillestrøm, but they identify effects the IGF should test rather than assume away.

Informal advantage becomes agenda power over time. Relationships formed onsite can produce next year's proposal, speaker list or coalition. Entities with regular travel support accumulate recognition. A remote attendee may supply a strong question without entering the network that carries it forward.

The solution is not to romanticize corridors or force online replicas of receptions. It is to create visible follow-up: opt-in introductions, mixed small groups, published contact routes, post-event meetings, open drafting spaces and response logs. Informal access cannot be equalized completely, but its institutional consequences can be reduced.

The remote majority occupied more endpoints. The onsite minority still occupied the densest shared environment.

Presence did not determine whose words entered the record

The Lillestrøm Messages synthesized discussion across 262 sessions. Session reports, transcripts, videos and summaries created a large public record. This archive lets remote entities follow material they missed and gives later institutions a navigable account.

Synthesis is a second agenda. Rapporteurs decide what a session report emphasizes. Editors identify recurring points. Annual messages compress disagreements into readable propositions. These acts are necessary and consequential.

A remote-majority meeting should ask whether remote contributions survived each stage. How many session reports included online interventions? Were remote questions attributed accurately? Did the annual synthesis draw from chat and hub reports or mainly from speakers and onsite discussion? Could a entity correct a material omission?

The official output page says the Messages capture views expressed through the 262 sessions. That status is properly narrower than a negotiated outcome. It should not be expanded into the voice of all 9,435-plus entities. Many did not attend every session, and no one voted on the final synthesis.

This is a classic route to consensus capture. A diverse crowd generates institutional prestige; a smaller set of organizers, speakers, rapporteurs and editors produces text; later readers hear the text as what "the community" decided. The problem is not synthesis. It is missing attribution and denominator.

Good messages identify the kind of evidence behind each point, retain material disagreement and link to session records. They avoid claiming that repetition equals authorization. Remote participation becomes meaningful when a remote contribution can be traced into the record, not when the total is printed on its cover.

A changed composition did not erase institutional inequality

Compared with Riyadh in 2024, the Lillestrøm meeting reported a major shift in stakeholder composition. Civil-society and technical-community shares increased while government and private-sector shares fell. Gender balance improved from roughly 31 per cent female in 2024 to about 46 per cent in 2025. These are important changes.

Location almost certainly contributed. Riyadh's entity population was heavily government and private sector, male, and based in Asia Pacific and the host region. Lillestrøm was heavily Western European and from developed countries, with civil society the largest group in the final report. A rotating forum can broaden itself over time precisely because each host produces different access.

Annual comparison does not prove that remote mode caused the shifts. Host geography, recruitment, dates, theme, institutional travel decisions and category definitions also changed. The official materials present percentages that are not entirely consistent. Causal claims need entity-level and year-over-year methods.

Nor does improved composition establish equal influence. Nadia Tjahja's 2025 research on meaningful IGF participation argues for moving from counting voices to examining how they are integrated. Earlier work with Trisha Meyer and Jamal Shahin showed the diversity hidden inside the civil-society label. A larger civil-society share says little about which kinds of organizations organized sessions or reached outputs.

The Lillestrøm result should be celebrated as broader entry and investigated as a distribution of roles. Composition is evidence. It is not the conclusion.

Remote entities could influence without receiving authority

The insistence that remote majority did not govern should not be mistaken for saying online participation was passive or irrelevant. A well-placed question can change a panel's direction. A remote speaker can supply evidence unavailable in the room. A hub can identify a regional consequence that later enters a report. A contact made online can lead to work in a national process.

Influence often travels without formal power. Brenden Nonnecke's study of the East Africa IGF found capacity building, knowledge sharing and policy transfer alongside persistent differences in institutional resources. That combination is more plausible than either extreme: that attendance changes nothing, or that all entities become equal governors.

The IGF is designed for this kind of influence. It can make issues visible, translate between institutions, support learning and carry recommendations toward bodies that can act. Remote access increases the pool of possible evidence and relationships.

The institutional duty is to trace the path. If an online intervention changes a session report, say so. If a ministry later cites the report, identify the ministry's authority and consultation. If an operator changes practice, credit the information but preserve the operator's responsibility. If nothing happens, do not convert participation into an outcome.

Remote entities need no fictional mandate for their contributions to matter. Evidence, experience and reasoning are enough to influence an open forum. Authority remains with the institution that is answerable for the later decision.

The right unit is an opportunity chain

A hybrid participation audit should follow each layer from entry to effect.

The first layer is access: registration, platform entry, hub access, language, bandwidth, time zone and accessibility. The second is agenda entry: thematic submissions, stocktaking comments, session proposals and consultation participation. The third is programme role: organizer, speaker, moderator, rapporteur or general entity.

The fourth is interaction: questions raised, questions recognized, speaking time, technical interruption and mixed-mode exchange. The fifth is record: inclusion in reports, transcripts, messages or intersessional work, with correction available. The sixth is network: opt-in contacts, working groups and follow-up participation. The seventh is institutional response: evidence that a body with authority considered, adopted, modified or rejected the input.

At each layer, compare onsite, individual online and hub participation. Do not assume a person must advance through every layer. Listening and learning can be valuable ends. The purpose is to prevent a count at the first layer from being advertised as success at all seven.

This opportunity chain would reveal where Lillestrøm succeeded. It likely expanded first-layer access dramatically. Official reports show some remote speakers, hubs and hybrid facilities, indicating movement into programme and interaction. Published evidence is weaker on comparative speaking time, report inclusion, network formation and downstream response.

The chain also identifies practical improvements. If online proposals succeed but questions are ignored, moderation needs attention. If questions are heard but no follow-up exists, record and handoff need attention. If hubs are active but absent from statistics, counting needs attention.

Better measurement turns hybrid design from a promise into an accountable institution.

Five reforms for a genuinely remote-shaped forum

First, publish mode across the full annual cycle. The Forum should distinguish onsite, individual online and hub contributors to thematic calls, MAG consultations, proposal teams, accepted sessions, speaking roles and stocktaking. Meeting-week registration alone is too late.

Second, make online intervention outcomes visible. Each session should report remote and onsite questions received, recognized and left unanswered, along with major technical failures and the disposition of written comments. This need not name individuals.

Third, design for time-zone rotation and asynchronous consequence. Repeat selected sessions, rotate major tracks across hours, accept pre-submitted interventions and require a response to material questions that cannot be heard live. A recording is access to content, not access to deliberation.

Fourth, create durable mixed-mode networks. Offer opt-in small groups, remote office hours with relevant institutions, post-meeting meetings and open contributor spaces. Ensure remote entities can join the preparation of future proposals rather than merely watch this year's sessions.

Fifth, reconcile the numbers. Define registration and participation, explain unique-person rules, separate stream views, identify hub counting and publish corrected stakeholder and regional tables. The difference between 8,833 registrations and more than 9,435 entities should not be left for readers to guess.

None of these reforms gives online attendees a vote over institutions they do not govern. They make the Forum's own agenda and knowledge functions more responsive to the majority it invited through screens.

Remote-first must begin before the broadcast

The phrase "remote-first" is often reduced to meeting etiquette: invite the online question first, test the microphone and keep a moderator in the virtual room. Those practices matter, but they begin after the most consequential choices. A genuinely remote-shaped forum would start with how issues are solicited, proposals are assembled and advisory meetings are scheduled.

Calls should be usable on low bandwidth and available early enough for organizations without dedicated conference staff. Proposal support should help new teams understand the format without steering their substance. Consultations should rotate time zones, accept structured written input and publish how each class of input affected the result. Evaluation should test whether an organizing team gives remote contributors authorship, not merely access.

Speaker confirmation is another upstream boundary. Requiring some onsite presence can protect sessions against connection failure and preserve room interaction, but it can also make travel capacity a condition of proposal viability. The institution should disclose when onsite requirements apply, provide exceptions and support, and assess whether remote specialists hold central speaking roles. A panel designed around onsite principals with one remote respondent is hybrid in transmission and physical in authority.

Remote-first scheduling would reserve time for mixed interaction before every speaker list becomes crowded. Reporting forms would contain separate fields for remote contributions and unanswered interventions. Follow-up dates would be announced with the session, giving people outside the venue a predictable route into later work.

The standard is not that remote entities must dominate. It is that location should not determine who can originate an issue, own a proposal, frame the evidence or continue the work. By the time a broadcast starts, much of that distribution has already been settled. Lillestrøm proved the Forum can bring a remote majority to the door. The next design question is whether remote participation begins while the house is still being drawn.

What Lillestrøm permits and does not permit us to conclude

The evidence permits several strong claims. Remote participation exceeded onsite participation by a large margin. Hybrid facilities enabled thousands to join without travelling. Thirty-three hubs extended collective access across 21 countries, especially in Africa. Civil-society and technical-community presence increased, and the overall gender distribution improved substantially from 2024.

The evidence also shows persistent concentration. Workshop places were scarce and selected through the MAG process months before the annual crowd assembled. The host controlled essential facilities and high-level convening conditions. Session organizers and moderators mediated who became visible. The onsite population retained unusually rich informal access. Most entities came from developed countries, and the host region dominated geography.

The evidence does not tell us that remote attendees were passive. It does not show that hosts or MAG members abused their roles. It does not show that every onsite entity had influence or every online entity lacked it. It does not support a clean causal claim that remote mode produced the changes in stakeholder or gender composition.

Above all, it does not support a mandate claim. The remote majority was not elected, population weighted or asked to approve common positions. Its presence improved the evidence environment. It did not create a remote sovereign.

Conclusion: access expanded faster than control changed

Lillestrøm was a milestone because the physical venue no longer contained most reported entities. More than six thousand people joined online, compared with 3,435 onsite. The Forum supported remote hubs, virtual meeting access, online speakers, moderators, captions, streams and digital spaces. The cost of entry fell for many people who could not have travelled to Norway.

That achievement changed the meeting's reach. It did not automatically change the location of agenda power. The MAG and Secretariat had already structured and selected the programme. Organizers had framed sessions and invited speakers. The host had supplied the material and high-level environment. Moderators governed the bridge between chat and room. Onsite entities moved through a dense network of unplanned encounters unavailable in the same form online.

The answer is not to diminish remote participation or abolish programme stewardship. Global meetings need selection, facilities, moderation and accountable editors. The answer is to disclose where each power sits and test whether remote access reaches it.

Future IGFs should be able to show that online contributors entered thematic calls, won programme places, organized sessions, received speaking time, influenced reports, formed continuing relationships and reached institutions with authority. They should publish mode-specific opportunity without pretending that every entity sought the same depth of involvement. They should correct denominator conflicts before turning totals into institutional claims.

Lillestrøm's remote majority proved that access can scale beyond a venue. It also exposed a harder truth: connection is the first mile of participation, not the last. A raised hand must still become speech. Speech must still enter a record. A record must still reach an institution. That institution must still answer under its own mandate.

The Forum became more open in 2025. Whether it became more distributed depends on those later links. Until they are measured, the honest verdict is precise: the online public grew faster than the architecture of agenda power changed.

Sources