Revised legacy report 03: Representing genuine human diversity in AI training and evaluation

Status: REWRITTEN October 11, 2026. This edition replaces, rather than reproduces, the former privacy-centered report. The original research was retained in the user's earlier ZIP, not in this current canonical archive. Consult revision_history/CORRECTIONS_AND_PROVENANCE.md for the reasons and full source mapping.

Editorial position and scope

Proposed mission: Cognitive liberty means that people can think and inquire freely, seek and receive information, voluntarily share their views and evidence, encounter disagreement, and challenge opaque decisions that prevent lawful information from being found. Access does not equal endorsement: freedom to inspect a claim is compatible with labeling its evidence weak, correcting falsehoods, and applying proportionate lawful restrictions. Individual privacy remains an option and a safeguard, not the central objective or a demand that people stop sharing with knowledge systems.

Status of this document: Independent desk research as of October 11, 2026. It uses public sources and the existing public website. It is not an internal site audit, a measured search-ranking experiment, an inspection of proprietary AI training datasets, or legal advice. Numbered [S##] citations resolve to SOURCES.md.

Revised conclusion

The public site is not an empty shell: it has an accessible Resource Navigator, organizations directory, standards directory and evidence ledger. The proper critique is that discovery paths still emphasize private inquiry, not that the linked pages do not resolve.

Diversity is both a dataset coverage problem and a model-behavior problem. Original human writing from less represented languages and perspectives is worth preserving, but publishing a viewpoint does not ensure its inclusion in a training run. A contributor-governed knowledge commons and targeted evaluation provide more controllable interventions than simply allowing unlimited scraping.

Findings and evidence

Finding 1. UNESCO calls for inclusive AI teams, training datasets and fair access, with specific attention to marginalized and multilingual communities. [S02][S06]

Finding 2. UNESCO’s multilingual roadmap identifies language technologies as part of broadening digital representation. [S25]

Finding 3. Poole-Dayan and colleagues reported in July 2026 that their audit of publicly available frontier-lab materials found no clear evidence that pluralism was a named training/evaluation objective in production systems. This is limited by publicly disclosed evidence. [S03]

Finding 4. Shumailov et al. studied degradation from recursively generated training data. Their “model collapse” result does not establish that suppressing unpopular ideas causes the same technical phenomenon. [S05]

Corrected analysis

The AI pipeline has at least six selection points: which people publish; what is crawlable; how content is licensed; how text is filtered and weighted; whose preference judgments count; and which evaluations a product must pass. Diversity can be lost at any stage, including after pretraining.

The highest-impact intervention may be evaluation, not simply dataset growth. If models fail to recognize minority philosophies or confidently misattribute dissenting claims, targeted evaluation can expose those errors even when the training corpus is undisclosed.

A knowledge commons should label contributions as eyewitness account, opinion, scholarly claim, research dataset, artistic expression or proposed hypothesis. Preserving dissent does not require laundering conjecture into fact.

Contribution permissions should distinguish searchable public display, retrieval and quotation, research dataset release, and training reuse. These are different decisions and may involve different legal or contractual terms. Contributor voice grows stronger when attribution and correction mechanisms exist.

Site architecture continuity

Rebuild the site map on the actual accessible directories. An evidence graph should contain Person, Organization, Work, Claim, Position, Critique, Evidence, Language and Source URL nodes. Each edge must carry a relationship label (“argues for,” “criticizes,” “contains evidence for”) and review date. No claim that /cognitive-liberty-resource-navigator, /research-evidence, /cognitive-liberty-organizations, or /cognitive-liberty-standards is inaccessible should remain. Audit their navigation and relevance rather than fabricating an outage. [S19][S21][S22][S23]

Design recommendation

Add entry points for “Share a source,” “Explore disagreement,” “Find underrepresented authors” and “Challenge a missing result” to the present navigation.

• Design an openly readable contributor registry with stable IDs, language tags, viewpoint descriptions and provenance links.
• Pilot five multilingual and multi-perspective topic collections, with independent editors and a published inclusion policy.
• Publish reproducible AI tests for accurate description of viewpoints, not merely agreement with them.

Proposed validation

• Coverage across languages and self-described perspectives rather than forced ideological categories.
• Model description accuracy for underrepresented viewpoints and false consensus frequency.
• Attributable contributions reused in publicly observable research or citation systems.

Limits and caveats

Open publishing cannot guarantee foundation-model training. Publishing data unconditionally can expose participants or undermine their choices, and more text is not automatically more representative.

Source references

• [S02] Recommendation on the Ethics of Artificial Intelligence — UNESCO, 2021: https://www.unesco.org/en/legal-affairs/recommendation-ethics-artificial-intelligence. Relevance: Calls for cultural and political diversity, participatory AI, freedom of expression and information.
• [S03] A Roadmap to Impactful Pluralistic Alignment Research — Poole-Dayan et al., 2026, preprint: https://arxiv.org/abs/2607.22305. Relevance: Audit of published behavior documents; reports no clear evidence of production pluralistic training/evaluation in scope, not proof of none.
• [S05] AI models collapse when trained on recursively generated data — Shumailov et al., Nature, 2024 (correction 2025): https://www.nature.com/articles/s41586-024-07566-y. Relevance: Study about recursive training on synthetic content; not synonymous with viewpoint censorship.
• [S06] AI systems and language diversity — UNESCO linguistic diversity coalition: https://www.unesco.org/ethics-ai/en/coalition-linguistic-diversity-artificial-intelligence. Relevance: Documented linguistic underrepresentation and need for broader language access.
• [S25] Global Roadmap for Multilingualism in the Digital Era — UNESCO, 2025/2026: https://www.unesco.org/en/global-roadmap-multilingualism. Relevance: Recommendations for language technologies and equitable representation.