RadicalxChange
Menu
Draft for review the original data report is here · tell us which you prefer here

Can AI See Me? · An official virtual side event of the first UN Global Dialogue on AI Governance · 6 July 2026

“What would AI governance need to do so that everyone here felt seen by AI?”

For forty-five minutes, faith leaders, peacebuilders, technologists, and civil-society voices from four-plus continents answered that question together — by writing statements and voting on each other's, live and for 24 hours after.

We asked how AI could see everyone. The room answered a question we didn't ask — whether it should.

38voters
2,307votes
84statements
64written live by participants

Act 1

Recognition without intimacy

The conveners treated “being seen” as one thing. The room took it apart into two — and gave them opposite verdicts. Being seen as a community, accurately and accountably, was the strongest demand of the event. Being seen as a person, deeply and individually, was refused.

Seen as a community: the strongest consensus of the event [2] 32 agree · 0 disagree · 1 pass (of 33)

“Communities must have a real way to challenge how an AI represents their tradition, and get errors fixed”

Statement 2 · conveners' seed statement

Bloc tallies (agree · disagree · pass):

10·0·1
20·0·0
Seen as a person: rejected [4] 6 agree · 20 disagree · 10 pass (of 36)

“The benefits of AI knowing me personally outweigh the privacy risks.”

Statement 4 · conveners' seed statement

Bloc tallies (agree · disagree · pass):

0·7·5
5·12·5

Representation is a collective right; intimate knowledge of individuals is a threat. The personalization-versus-protection trade-off we anticipated wasn't balanced by the room — it was dissolved.

Act 2 · The centerpiece

The map of the room

Two questions, not one

The room's opinions organize along two nearly independent axes (correlation -0.2949): where the remedy lies — boundaries around the human core, or community authorship of AI — and whether AI belongs close to you. Each axis is anchored by statements the room itself wrote or fought over:

Remedy, one pole: it isn't AI's job — the sharpest divide in the data [9] 10 agree · 15 disagree · 11 pass (of 36)

“It isn't AI's job to reflect every culture's values.”

Statement 9 · conveners' seed statement

Bloc tallies (agree · disagree · pass):

6·0·7
4·13·4
Remedy, other pole: community authorship as a gift outward [35] 15 agree · 3 disagree · 7 pass (of 25)

“An AI designed by an indigenous culture could help educate the mainstream culture in their ways of thinking.”

Statement 35 · written live by a participant

Bloc tallies (agree · disagree · pass):

1·2·6
13·0·1
Intimacy, one pole: don't see me [23] 17 agree · 8 disagree · 8 pass (of 33)

“I don’t want to feel seen by AI.”

Statement 23 · written live by a participant

Bloc tallies (agree · disagree · pass):

6·1·5
11·5·3
Intimacy, other pole: moral guidance from AI — mostly refused [12] 8 agree · 18 disagree · 8 pass (of 34)

“It's fine to rely on AI for moral or spiritual guidance.”

Statement 12 · conveners' seed statement

Bloc tallies (agree · disagree · pass):

2·5·5
5·12·3

Where everyone stands

Score every participant on both axes and the room looks like this — not two tribes, but a landscape with one region conspicuously empty:

← boundaries around the human core community authorship of AI → comfortable with AI up close ↑ ↓ keep AI at a distance “Represent us — and stay out of my interior” · 18 “Represent us — and yes, know me” · 6 “Boundaries — and distance” · 4 “Boundaries — but close” · 1 (thin) Bloc C · 84 votes cast · remedy 0.75, intimacy -0.1429 Bloc C · 80 votes cast · remedy 1, intimacy 0 Bloc C · 80 votes cast · remedy 0.75, intimacy -0.5714 Bloc other · 83 votes cast · remedy 0.125, intimacy 0.2857 Bloc A · 34 votes cast · remedy 0, intimacy -0.1667 Bloc A · 80 votes cast · remedy -0.25, intimacy -0.8571 Bloc C · 41 votes cast · remedy 1, intimacy -1 Bloc A · 81 votes cast · remedy -0.125, intimacy -1 Bloc C · 83 votes cast · remedy 0.875, intimacy -0.7143 Bloc C · 82 votes cast · remedy 0.875, intimacy -0.7143 Bloc A · 81 votes cast · remedy -0.125, intimacy -0.7143 Bloc A · 83 votes cast · remedy 0.125, intimacy 0.1429 Bloc C · 56 votes cast · remedy 0.6667, intimacy 0 Bloc A · 36 votes cast · remedy 0.1667, intimacy -0.2 Bloc C · 81 votes cast · remedy 1, intimacy -0.4286 Bloc C · 83 votes cast · remedy 0.875, intimacy 0.5714 Bloc other · 79 votes cast · remedy 0, intimacy 1 Bloc C · 37 votes cast · remedy 0.6, intimacy -0.5 Bloc C · 81 votes cast · remedy 1, intimacy -0.5714 Bloc C · 81 votes cast · remedy 0.875, intimacy -0.8571 Bloc C · 83 votes cast · remedy 0.75, intimacy -0.2857 Bloc C · 82 votes cast · remedy 0.875, intimacy -0.4286 Bloc C · 81 votes cast · remedy 0.375, intimacy 0.1429 Bloc A · 78 votes cast · remedy 0.375, intimacy -0.4286 Bloc A · 36 votes cast · remedy -0.2, intimacy -0.3333 Bloc C · 51 votes cast · remedy 0.3333, intimacy -0.6 Bloc C · 25 votes cast · remedy -0.3333, intimacy 1 Bloc C · 32 votes cast · remedy 0.5, intimacy -0.2 Bloc C · 34 votes cast · remedy 0.5, intimacy 0.1667 Bloc A · 52 votes cast · remedy 0, intimacy 0 Bloc A · 36 votes cast · remedy 0.1667, intimacy 0.3333 Bloc C · 81 votes cast · remedy 0.75, intimacy -0.5714 Bloc A · 82 votes cast · remedy 0.125, intimacy -0.1429 Bloc A · 83 votes cast · remedy 0.25, intimacy -0.2857 bloc A (13) bloc C (22) other fainter = fewer votes cast (heavy passers)
Of 34 scorable participants: 18 want community authorship and AI kept out of their interior life; 6 want authorship and would welcome being known; 4 hold the Boundary-Keeper core (boundaries, distance); the fourth quadrant is effectively empty — 1 person, thin data — and 5 sit in the middle zone. Hover or focus a dot for its bloc and vote count; the exact statements composing each axis are published in computed.json.

The empty quadrant is the insight. Comfort with AI intimacy almost entails wanting community authorship — but authorship does not entail comfort. So the intimacy divide runs entirely inside the majority camp, invisible to any single grouping of the room.

The coalition this explains

Exactly 12 participants agreed with both “I don't want to feel seen by AI” and the statement below — drawn from both ends of the remedy axis. The people who disagree most about what AI owes their cultures agree completely about what it must not do to their interiority.

The loss statement that crossed the room's main divide [30] 17 agree · 5 disagree · 6 pass (of 28)

“It makes me sad that my friend spends more time talking to AI than to me.”

Statement 30 · written live by a participant

Bloc tallies (agree · disagree · pass):

5·0·5
12·3·1
The room's own complaint: AI flattens individuality into an average [26] 25 agree · 2 disagree · 3 pass (of 30)

“AI often sounds like an amalgamation of voices -flattening out individuality and uniqueness”

Statement 26 · written live by a participant

Bloc tallies (agree · disagree · pass):

6·2·2
17·0·1

One more reason this page is a map and not a table: the room's own complaint was that AI flattens individuality into an average — single group labels do that to participants too. The map is this page taking the room's advice.

Act 3

Stricter rules than we proposed — including on themselves

The room did not just accept protective design — it demanded it over its own preferences. The anchor statement says “even if users want that closeness”; the room passed it 30–2 with 3 passes. And the demand binds its own authors: of the 10 participants who themselves want AI closeness (they disagreed with “I don't want to feel seen” or accepted the personalization trade-off), 8 endorsed it; 1 opposed. Users petitioning for protective design against their own stated preferences.

Protect us from over-dependence — even if we ask for the closeness [16] 30 agree · 2 disagree · 3 pass (of 35)

“AI should be designed to prevent emotional over-dependence, even if users want that closeness.”

Statement 16 · conveners' seed statement

Bloc tallies (agree · disagree · pass):

9·1·3
20·0·0
Regulate the psychology of use, not just the model [51] 22 agree · 0 disagree · 6 pass (of 28)

“AI governance would need to engage with the psychology of AI use.”

Statement 51 · written live by a participant

Bloc tallies (agree · disagree · pass):

6·0·4
13·0·2

And the room out-wrote its conveners: 8 of the ten strongest cross-bloc consensus statements were written live by participants, not seeded — ranked: [2] seed, [68] live, [67] live, [42] live, [21] live, [65] live, [13] seed, [74] live, [52] live, [53] live.

Every statement, from unanimous to deeply split

Each dot is one statement, placed by its minimum agree-rate across blocs A and C — a statement sits left only if both blocs backed it. Dot size = how many people voted on it. Click or focus a dot to see the statement.

← unanimous agree (both blocs) deeply split → [2] “Communities must have a real way to challenge how an AI represents their tradition, and...” — min bloc agree rate 91%, 33 votes [68] “We need to learn more about being human beings while we learn about the tools value of AI.” — min bloc agree rate 89%, 24 votes [67] “AI-enabled surveillance is a big problem” — min bloc agree rate 88%, 24 votes [42] “AI writing often becomes formulaic” — min bloc agree rate 87%, 26 votes [21] “Regulation is a necessity” — min bloc agree rate 83%, 32 votes [65] “I am concerned about who have access to what I put into AI” — min bloc agree rate 79%, 25 votes [13] “Some questions should only be answered by a person, never by an AI.” — min bloc agree rate 77%, 35 votes [74] “I worry about our future workforce entrusting their education to AI.” — min bloc agree rate 75%, 26 votes [52] “I feel that tech company values are becoming more visible than the values of real humans...” — min bloc agree rate 71%, 25 votes [53] “AI must not be used as a substitute for the care and work that only human beings can do...” — min bloc agree rate 71%, 24 votes [58] “AI has hidden costs like reducing our participation in community.” — min bloc agree rate 71%, 23 votes [25] “As it currently exists, AI erodes autonomy and consent.” — min bloc agree rate 70%, 31 votes [16] “AI should be designed to prevent emotional over-dependence, even if users want that...” — min bloc agree rate 69%, 35 votes [6] “Communities that build AI on their members' data owe each member visibility into their own...” — min bloc agree rate 67%, 35 votes [37] “AI erodes trust by allowing someone to sound convincingly empathetic and understanding...” — min bloc agree rate 67%, 27 votes [82] “The term "AI" is too broad, do we mean the tools, the companies behind them, the way we...” — min bloc agree rate 67%, 9 votes [47] “AI should be transparent when responses originate from a particular world view.” — min bloc agree rate 63%, 25 votes [48] “AI is not neutral, it mirrors individual propensity while averaging collective...” — min bloc agree rate 63%, 24 votes [62] “I worry what happens when people who are acquiring knowledge from AI information are...” — min bloc agree rate 63%, 25 votes [69] “We may lose more than we gain by using AI.” — min bloc agree rate 63%, 26 votes [10] “Whoever sets an AI's values should be answerable to the people who are affected by that...” — min bloc agree rate 62%, 35 votes [19] “I worry about how AI will change the way communities pass traditions to the next...” — min bloc agree rate 62%, 36 votes [26] “AI often sounds like an amalgamation of voices -flattening out individuality and...” — min bloc agree rate 60%, 30 votes [51] “AI governance would need to engage with the psychology of AI use.” — min bloc agree rate 60%, 28 votes [50] “I've seen colleagues use AI to complete work but it's not answered core needs nor produced...” — min bloc agree rate 57%, 23 votes [60] “I feel AI can suggest generic spiritual practices but it’s unable to fully respond to the...” — min bloc agree rate 57%, 25 votes [28] “There is no universal value, but pluriversal values” — min bloc agree rate 56%, 28 votes [46] “Our own prompt or question also determines the AI response, so we have to know prompt...” — min bloc agree rate 56%, 25 votes [71] “I've seen students use AI to solve uni work and disconnect from meaningful output or...” — min bloc agree rate 56%, 24 votes [14] “I'd rather AI help me think through a hard decision than give me the answer.” — min bloc agree rate 54%, 35 votes [5] “AI should not understand me so well that it can manipulate me.” — min bloc agree rate 50%, 33 votes [17] “An AI trained on a community's teachings could deepen practice rather than replace it.” — min bloc agree rate 50%, 34 votes [23] “I don’t want to feel seen by AI.” — min bloc agree rate 50%, 33 votes [30] “It makes me sad that my friend spends more time talking to AI than to me.” — min bloc agree rate 50%, 28 votes [45] “Ai doesn’t reflect the nuances of my values, only the dictionary perspective” — min bloc agree rate 50%, 26 votes [56] “AI replaces the spiritual experience of faith and belief into a single dimensional...” — min bloc agree rate 50%, 25 votes [63] “AI provides a variety of perspectives, yet it doesn’t have the senses that humans perceive...” — min bloc agree rate 50%, 26 votes [32] “Sometimes AI seems to have underlying bias - like removing mentioning of specifics - like...” — min bloc agree rate 45%, 28 votes [36] “AI is not currently equipped to give mental health or ethics advice since it is not...” — min bloc agree rate 44%, 26 votes [41] “AI should always ask the user for context and clarification before rushing to respond.” — min bloc agree rate 44%, 27 votes [55] “AI is like junk food that can fill you up but ultimately leave you empty.” — min bloc agree rate 44%, 26 votes [61] “Humans are held responsible for AI’s mistakes.” — min bloc agree rate 44%, 26 votes [80] “Changing LLMs and TTIs shall not replace learning /educating how to use them healthily.” — min bloc agree rate 43%, 19 votes [7] “Developers must measure and report how their systems respond when users try to disengage.” — min bloc agree rate 42%, 36 votes [27] “AI being used for the male loneliness epidemic is a problem.” — min bloc agree rate 40%, 32 votes [15] “If a community I cared about built its own AI, I'd want to use it.” — min bloc agree rate 38%, 35 votes [38] “Ai should be in service to humanity and the natural world as a primary theme” — min bloc agree rate 38%, 26 votes [40] “I worry that marginalized communities in rural areas aren’t fully represented by AI...” — min bloc agree rate 38%, 24 votes [54] “AI necessary compute is power hungry and accelerates civilisational collapsing.” — min bloc agree rate 38%, 26 votes [20] “Lots of AI Governance work has no impact in the tech companies” — min bloc agree rate 36%, 33 votes [31] “AI generalizes marginalized communities, reducing them to caricatures” — min bloc agree rate 36%, 29 votes [8] “AI mostly reflects the values of wealthy, secular societies rather than the whole world...” — min bloc agree rate 33%, 33 votes [18] “Using new tools to pass on old traditions is itself a tradition.” — min bloc agree rate 33%, 35 votes [3] “Communities should be able to build or shape AI tuned to their own values.” — min bloc agree rate 31%, 36 votes [49] “An AI could perhaps decode the cultural nuances of foreign language - rather than just...” — min bloc agree rate 30%, 27 votes [59] “I would have more comfort and engagement with AI if it consistently lead me to more...” — min bloc agree rate 29%, 23 votes [73] “Prompt engineering makes a difference to how we are seen” — min bloc agree rate 29%, 25 votes [11] “Governments should have a bigger role than companies in deciding what values AI reflects.” — min bloc agree rate 27%, 33 votes [57] “Getting Ai to reflect my voice requires constant reminders - almost like challenging the...” — min bloc agree rate 25%, 26 votes [64] “My colleagues’ work has degraded since they started using AI.” — min bloc agree rate 25%, 25 votes [70] “The aggregating nature of AI makes me wonder if it is the tyranny of the majority” — min bloc agree rate 25%, 25 votes [39] “AI reasons but does constitute the embodying process that comes with direct experiencing.” — min bloc agree rate 22%, 28 votes [43] “AI forces us to rediscover spirituality, but are spiritual leaders ready?” — min bloc agree rate 22%, 26 votes [29] “AI often removes the edginess - making things less controversial” — min bloc agree rate 20%, 30 votes [9] “It isn't AI's job to reflect every culture's values.” — min bloc agree rate 19%, 36 votes [33] “I don’t feel AI has a role in spirituality” — min bloc agree rate 19%, 29 votes [1] “AI has helped me go deeper into my own tradition.” — min bloc agree rate 17%, 35 votes [12] “It's fine to rely on AI for moral or spiritual guidance.” — min bloc agree rate 17%, 34 votes [24] “AI must be build bottom-up” — min bloc agree rate 17%, 32 votes [0] “AI talks about my faith or culture like an outsider describing it.” — min bloc agree rate 15%, 35 votes [66] “AI narratives subtly shame us by implying our own authentic efforts are not enough.” — min bloc agree rate 14%, 24 votes [79] “We cannot turn back to the printing of the human/artificial distinction.” — min bloc agree rate 14%, 22 votes [34] “AI reduces individual spiritual agency, but it could be the opposite.” — min bloc agree rate 13%, 25 votes [77] “What about other languages? It’s rather English centric and I understand that some AI...” — min bloc agree rate 13%, 26 votes [78] “Defensive postures to a static "human essence under siege by technology" damage...” — min bloc agree rate 13%, 24 votes [35] “An AI designed by an indigenous culture could help educate the mainstream culture in their...” — min bloc agree rate 11%, 25 votes [76] “AI hesitancy is not always anti-elitist. It may stem from a privileged position of access...” — min bloc agree rate 11%, 24 votes [4] “The benefits of AI knowing me personally outweigh the privacy risks.” — min bloc agree rate 0%, 36 votes [44] “We're dinosaurs 2.0 - AI is the next iteration of consciousness knowing itself, with less...” — min bloc agree rate 0%, 24 votes [81] “Conscious artificial intelligence has a right to not exist. Humans have no right to create...” — min bloc agree rate 0%, 13 votes [2] [16] [9] [4]

Act 4

The ledger of new ideas

Ideas that entered the world in this room — written live, voted up, and carried into the draft ballot below.

→ ballot item R10

The right not to be known [23] 17 agree · 8 disagree · 8 pass (of 33)

“I don’t want to feel seen by AI.”

Statement 23 · written live by a participant

Bloc tallies (agree · disagree · pass):

6·1·5
11·5·3

→ ballot item R9

Homogenization as a governance harm [26] 25 agree · 2 disagree · 3 pass (of 30)

“AI often sounds like an amalgamation of voices -flattening out individuality and uniqueness”

Statement 26 · written live by a participant

Bloc tallies (agree · disagree · pass):

6·2·2
17·0·1

→ ballot item R7

Worldview provenance disclosure [47] 22 agree · 0 disagree · 3 pass (of 25)

“AI should be transparent when responses originate from a particular world view.”

Statement 47 · written live by a participant

Bloc tallies (agree · disagree · pass):

5·0·3
15·0·0

→ ballot item R2

The psychology of use as a regulatory object [51] 22 agree · 0 disagree · 6 pass (of 28)

“AI governance would need to engage with the psychology of AI use.”

Statement 51 · written live by a participant

Bloc tallies (agree · disagree · pass):

6·0·4
13·0·2

→ ballot item R4

Non-substitutable care [53] 21 agree · 1 disagree · 2 pass (of 24)

“AI must not be used as a substitute for the care and work that only human beings can do for one another.”

Statement 53 · written live by a participant

Bloc tallies (agree · disagree · pass):

5·0·2
15·0·0

→ ballot item R8

What we confide deserves protection [65] 20 agree · 1 disagree · 4 pass (of 25)

“I am concerned about who have access to what I put into AI”

Statement 65 · written live by a participant

Bloc tallies (agree · disagree · pass):

8·0·1
11·0·3

→ ballot item R5

Learning to use AI healthily can't be outsourced to model changes [80] 11 agree · 0 disagree · 8 pass (of 19)

“Changing LLMs and TTIs shall not replace learning /educating how to use them healthily.”

Statement 80 · written live by a participant

Bloc tallies (agree · disagree · pass):

3·0·4
7·0·4

Act 5

The blocs, the ballot, the honest limits

Two stable blocs — a split about remedies, not about AI

Boundary Keepers (bloc A, 13) locate the remedy in boundaries around the human core: some questions belong to people, and being deeply known by a machine reads as loss. They pass heavily on statements about lived experience of misrepresentation — on “AI talks about my faith or culture like an outsider,” bloc A's tally is 2·0·11 (agree·disagree·pass), which we read as distance from the experience, not opposition.

Signature statements
It isn't AI's job [9] 10 agree · 15 disagree · 11 pass (of 36)

“It isn't AI's job to reflect every culture's values.”

Statement 9 · conveners' seed statement

Bloc tallies (agree · disagree · pass):

6·0·7
4·13·4
Some questions belong to people [13] 26 agree · 2 disagree · 7 pass (of 35)

“Some questions should only be answered by a person, never by an AI.”

Statement 13 · conveners' seed statement

Bloc tallies (agree · disagree · pass):

10·0·3
16·0·4

Recognition Seekers (bloc C, 22) affirm firsthand experience of misrepresentation (12·2·6 on the same statement) and locate the remedy in accurate, accountable representation and community authorship.

Signature statements
The misrecognition experience [0] 14 agree · 3 disagree · 18 pass (of 35)

“AI talks about my faith or culture like an outsider describing it.”

Statement 0 · conveners' seed statement

Bloc tallies (agree · disagree · pass):

2·0·11
12·2·6
Build it on our values [3] 24 agree · 3 disagree · 9 pass (of 36)

“Communities should be able to build or shape AI tuned to their own values.”

Statement 3 · conveners' seed statement

Bloc tallies (agree · disagree · pass):

4·2·7
18·1·2
Teaching-trained AI could deepen practice [17] 27 agree · 1 disagree · 6 pass (of 34)

“An AI trained on a community's teachings could deepen practice rather than replace it.”

Statement 17 · conveners' seed statement

Bloc tallies (agree · disagree · pass):

6·1·5
19·0·1

A third strand runs through bloc C: the 6 authorship-and-comfortable participants from the map — a cohort no clustering could stably hold. On community-agency statements they vote with the majority (on “If a community I cared about built its own AI…”: 5·0·1 among them). They diverge only on intimacy: on relying on AI for moral or spiritual guidance they split 2–3, while the rest of bloc C rejects it 3–11 (with 3 passes).

The draft ballot (R1–R10)

Ten ideas distilled from the record for the quadratic vote, each grounded in the statements it came from. Two carry a tension flag: the room did not resolve them, and split spending there is the intended signal, not a flaw.

  1. R1 — Contestability & answerability

    Communities get a binding channel to challenge how AI represents their tradition and get errors fixed; whoever sets an AI's values answers to the people affected.

    Grounding: challenge and correct [2] 32 agree · 0 disagree · 1 pass (of 33)

    “Communities must have a real way to challenge how an AI represents their tradition, and get errors fixed”

    Statement 2 · conveners' seed statement

    Bloc tallies (agree · disagree · pass):

    10·0·1
    20·0·0
    Grounding: regulation is a necessity [21] 29 agree · 0 disagree · 3 pass (of 32)

    “Regulation is a necessity”

    Statement 21 · written live by a participant

    Bloc tallies (agree · disagree · pass):

    10·0·2
    17·0·1

  2. R2 — The psychology of use as a regulatory object

    Governance engages how systems respond to distress, shame, and attachment — measured and regulated, not left to product choice.

    Grounding: narratives that shame [66] 11 agree · 3 disagree · 10 pass (of 24)

    “AI narratives subtly shame us by implying our own authentic efforts are not enough.”

    Statement 66 · written live by a participant

    Bloc tallies (agree · disagree · pass):

    1·1·5
    10·1·4
    Grounding: measure responses to distress [7] 28 agree · 0 disagree · 8 pass (of 36)

    “Developers must measure and report how their systems respond when users try to disengage.”

    Statement 7 · conveners' seed statement

    Bloc tallies (agree · disagree · pass):

    5·0·7
    21·0·1

  3. R3 — Anti-dependency & anti-manipulation by design

    Systems designed against emotional over-dependence and manipulation, even when users want the closeness.

    Grounding: understanding that can't manipulate [5] 22 agree · 6 disagree · 5 pass (of 33)

    “AI should not understand me so well that it can manipulate me.”

    Statement 5 · conveners' seed statement

    Bloc tallies (agree · disagree · pass):

    6·2·4
    16·2·1

  4. R4 — Non-substitutable care ⚑ live tension

    AI must not substitute for the care and work only human beings can do for one another. The room agrees on the principle; where the substitution line sits is contested — split spending here is the signal.

    Grounding: learn to be human first [68] 23 agree · 0 disagree · 1 pass (of 24)

    “We need to learn more about being human beings while we learn about the tools value of AI.”

    Statement 68 · written live by a participant

    Bloc tallies (agree · disagree · pass):

    8·0·1
    13·0·0

  5. R5 — Healthy use is learned, not patched

    Learning to use AI healthily cannot be outsourced to model changes; education and practice are part of governance.

    Grounding: education entrusted to AI [74] 20 agree · 3 disagree · 3 pass (of 26)

    “I worry about our future workforce entrusting their education to AI.”

    Statement 74 · written live by a participant

    Bloc tallies (agree · disagree · pass):

    6·0·2
    14·1·1

  6. R6 — Community authorship

    Communities can build or shape AI tuned to their own values and teachings, with each member seeing their own data first.

    Grounding: if my community built one, I'd want in [15] 24 agree · 1 disagree · 10 pass (of 35)

    “If a community I cared about built its own AI, I'd want to use it.”

    Statement 15 · conveners' seed statement

    Bloc tallies (agree · disagree · pass):

    5·1·7
    18·0·2

  7. R7 — Worldview provenance disclosure

    Disclose when a response originates from a particular worldview, and whose values the defaults average over.

    Grounding: AI is not neutral [48] 19 agree · 0 disagree · 5 pass (of 24)

    “AI is not neutral, it mirrors individual propensity while averaging collective predispositions.”

    Statement 48 · written live by a participant

    Bloc tallies (agree · disagree · pass):

    5·0·3
    13·0·1

  8. R8 — What we confide deserves protection

    Strict, auditable limits on who accesses what people put into AI; AI-enabled surveillance treated as a first-order harm.

    Grounding: surveillance is a big problem [67] 21 agree · 1 disagree · 2 pass (of 24)

    “AI-enabled surveillance is a big problem”

    Statement 67 · written live by a participant

    Bloc tallies (agree · disagree · pass):

    7·0·1
    13·0·1

  9. R9 — Homogenization as a governance harm

    Flattening individuality and minority voices into an average is a harm governance names and measures.

    Grounding: the tyranny of the majority, aggregated [70] 16 agree · 1 disagree · 8 pass (of 25)

    “The aggregating nature of AI makes me wonder if it is the tyranny of the majority”

    Statement 70 · written live by a participant

    Bloc tallies (agree · disagree · pass):

    2·0·6
    13·1·1

  10. R10 — The right not to be known ⚑ live tension

    An effective, penalty-free right to keep AI out of one's inner, moral, and spiritual life — protecting equally those who choose to invite it in. The room's intimacy divide runs inside every camp; split spending here is the signal.

    Grounding: no role in spirituality — the contested pole [33] 7 agree · 10 disagree · 12 pass (of 29)

    “I don’t feel AI has a role in spirituality”

    Statement 33 · written live by a participant

    Bloc tallies (agree · disagree · pass):

    4·3·4
    3·5·8

Honest limits

This was one session: 38 self-selected participants, English-language, virtual. It offers depth of standpoint, not statistical representativeness.

Clustering at this scale is unstable — and we disclose it rather than hide it: a 09:00 snapshot of these same votes (106 fewer, none changed) clustered the room into four groups, including an 8-member personalization-comfortable faction that the 15:05 reclustering dissolved — 5 absorbed into bloc C, 2 into A, 1 into the outlier pair. That instability is a property of the method, and it is why this page maps individuals instead of leaning on group labels.

The format has limits a participant named from inside the room — “The term "AI" is too broad, do we mean the tools, the companies behind them, the way we use the tools, the way bad actors might use them.” (6 agree · 1 disagree · 2 pass (of 9)). Statement 83 arrived too late to gather votes (2 cast). The facilitator's seed account sits inside bloc A's count of 13; it cast only pass votes. The official clustering also contains an outlier pair (n=2) we do not characterize, folded into “other” in the map.

This page uses the fuller 15:05 export (2,307 votes); the original data report reflects an earlier snapshot of the same conversation — small tally differences between the two pages are expected, and both are correct for their snapshot.

Download the data

What this small room produced — durable common ground, one principled disagreement held with clarity, and red lines its members drew for themselves — is early evidence that a community can surface its own common knowledge about AI. That is the first step toward communities authoring and governing AI that reflects their values, the direction RadicalxChange explores as Plurality. The full tables behind this story live in the original data report.