Fundamental Analysis: Humanity's Greatest Problems — and an LLM as a Living Constitution? (Claude)

Prompt 1: "What are the greatest problems of our time for people and humanity, and what would the best solution look like? Research the web and current scientific studies." Prompt 2: an examination of administrative logics, the debt brake, cross-domain costs, and whether an LLM-assisted constitutional authority could represent the interests of people not yet born.

This is the foundational analysis behind the articles “Three AIs, One Question”, “The Position of the Unborn” and the commentary “Only the Contribution Rate is Stabilized”. Two additional models received the same prompts independently (see replications).

Analysis 1: The Greatest Problems of Our Time

Core Finding: Research converges on six major problem clusters — and on a remarkable diagnosis: The true meta-problem is not a lack of knowledge or technology, but a failure of cooperation and governance under conditions of mutually reinforcing crises (“polycrisis”).

The Six Clusters

1. Ecological Destabilization. For the first time, seven out of nine planetary boundaries have been breached, ocean acidification joining as the latest (Planetary Health Check 2025, PIK); all trends are deteriorating. In 2024, global temperatures exceeded 1.5 °C for the first time. The Lancet Countdown 2025 (128 experts, 71 institutions) reports 12 of 20 health indicators at record levels and ~546,000 heat deaths per year (+63% since the 1990s).
2. Geopolitical Confrontation and War. The WEF Global Risks Report 2026 ranks geo-economic confrontation as the number one two-year risk. Doubly dangerous: directly (war, nuclear risk) and indirectly — it blocks the cooperation that all other clusters require.
3. Erosion of Democracy and Information Ecosystem. According to the V-Dem Democracy Report 2026, global democracy has fallen back to 1978 levels; 74% of the world population lives in autocracies, 44 countries are autocratizing — including western democracies for the first time. Freedom of expression is the element most frequently attacked first. The WEF ranks disinformation and polarization in positions 2–3.
4. Existential Risks: AI, Engineered Pandemics, Nuclear Weapons. Expert estimates vary widely (RAND 2025, P(doom) Survey), but the expected value is sufficiently high that leading AI researchers put AI extinction risks on par with pandemics and nuclear war. In the WEF long-term ranking, adverse AI effects made the largest jump of all risks (rank 30 → 5).
5. Stagnating Fight Against Poverty, Hunger, Disease. UN SDG Report 2025: only 35% of targets on track, 18% regressing; over 800 million people in extreme poverty. Antimicrobial resistance: 39 million direct deaths projected by 2050 (GRAM/Lancet).
6. The Inner Crisis: Loneliness. The WHO Commission on Social Connection (2025): 1 in 6 people is lonely, linked to ~871,000 deaths per year — adolescents are most affected (21%), not the elderly.

Connecting Framework: Polycrisis research shows: These crises are causally intertwined and reinforce each other in cascading fashion (climate stress → migration → polarization → autocratization → cooperation bottleneck → further climate stress). Treating them individually means losing.

The Best Solution

The evidence is clear about what works — the implementation deficit is the problem: Effective climate policy consists of policy mixes (pricing + regulation + subsidies; Stechemesser et al., Science 2024: only 63 out of 1,500 evaluated policies achieved major reductions). Halving emissions by 2030 would be feasible for under $100/t according to IPCC AR6. 92 million resistance deaths could be avoided simply through better access to existing antibiotics. Condensed into a single solution: restoring the capacity for cooperation itself — functioning democratic institutions, an intact information ecosystem, international coordination mechanisms. Nowhere is knowledge lacking; everywhere the collective capacity for execution is missing — and precisely this capacity is systematically destroyed by polarization, disinformation, autocratization, and geopolitical rivalry.

Analysis 2: Administrative Logics, Debt Brake, LLM as a Living Constitution

Contextualizing the Diagnosis

The thought experiment (“If I were not yet born…”) is precisely Rawls’ original position behind the veil of ignorance, radicalized across the intergenerational dimension. The administrative diagnosis has a name in academic literature: wrong pockets problem — whoever invests in prevention is not the one who captures the savings; rational silos systematically underinvest. The debt brake itself was justified with intergenerational justice; its structural flaw is the lack of separation between consumption and investment expenditure — it shields future generations from debt while surrendering them to investment backlogs. The March 2025 reform (special funds, defense exception) bypassed this selectively, rather than solving it structurally.

Institutions with This Mandate Already Exist — Proving the Gap

Wales (Well-being of Future Generations Act 2015, Commissioner), Japan (Saijo’s “Future Design”: Citizens as “imaginary future persons”; experiments show more than double the rate of future-oriented investments), Hungary (Ombudsman), UN (Declaration on Future Generations 2024). The pattern: The mechanism works, but it is localized, tied to individuals, voluntary — human future advocates tire, rotate, are politically curtailed, or remain toothless. The stance is taken selectively, never permanently.

Can an LLM Be the Living Constitution? Yes — as an Auditor, Not a Sovereign

Arguments For: Consistency is the core competency of an LLM panel (no re-election incentive, no departmental bias, no fiscal budget). The building blocks exist: The Habermas Machine (Tessler et al., Science 2024) demonstrated with over 5,700 participants that LLM mediation produces group statements with higher consensus than human mediators — while better integrating minority positions. C3AI (WWW 2025) shows how explicit constitutional principles are formulated and audited. And an auditing agent testing every law against the unborn position naturally calculates across departments — fiscal silos are not a constraint to it.

Counterarguments — and Design Answers: (1) Legitimacy: “good life” is contentious → the charter is explicit, versioned, and democratically enacted; the LLM merely executes. (2) Silent constitutional drift via model updates → multi-model panels, adversarial reviews, interpretative regression tests. (3) Hallucinated impact assessments → the LLM does not calculate autonomously; the quantitative backbone consists of validated models (GESIM) with sealed forecasts and backtests. (4) Degree of binding force: not a veto, but comply-or-explain — the legislature may deviate, but must publicly justify doing so. This shifts the payoff matrix without touching sovereignty — in the vocabulary of the Kantian Peace Machine: an additional institutional weight.

The Architecture: Three Existing Layers + One Missing

Layer	Foundation	Contribution
Normative	AI Dividend (Preprint 2026)	Rawls/Ostrom foundation, position of the unborn, convergence architecture
Institutional	Kantian Peace Machine (Preprint 2026)	Mechanism design: incentive architecture over moral appeals
Empirical	GESIM (Prototype)	Cross-domain cost cascades, sealed forecasts, validation ladder
Governance (Missing)	—	Living constitutional authority: Charter + LLM panel + comply-or-explain + quantitative backend

This fourth layer has been elaborated as an independent research project since July 10, 2026 (Working title: “The Position of the Unborn: LLM-Supported Constitutional Authorities as a Mechanism of Intergenerational Representation”).

---

Transparency: Analysis prepared on July 10, 2026 by Claude (Anthropic, Fable 5) with web research; editorially curated by LG. The unabridged draft resides in the editorial archive. Two independent parallel analyses: ChatGPT Sol (linked), Gemini (in progress).