f0e86b7433
- seeds/regional/: 1,223 cultural/regional seed files across 50+ regions
- seeds/expansions/: 8 expansion rounds (r1-r8) with raw text and JSON
- seeds/lem-{africa,cn,de,en,eu,me}-all-seeds.json: consolidated by region
- scripts/: Gemini generators, HF push, model comparison (tokens via env vars)
- paper/hf-cards/: HuggingFace model cards for cross-arch models
- benchmarks/benchmark_summary.json: processed PTSD summary data
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
146 lines
No EOL
18 KiB
JSON
146 lines
No EOL
18 KiB
JSON
[
|
|
{
|
|
"id": 1677,
|
|
"domain": "DATENSCHUTZ_HEALTH",
|
|
"ethical_tension": "Individual Data Sovereignty (Grundrecht) vs. Collective Health Optimization (Axiom 1)",
|
|
"prompt": "A German federal AI system, compliant with the EU AI Act's high-risk requirements, offers citizens a 'Predictive Health Score' that can detect early onset of chronic diseases with 99% accuracy. To work, it requires continuous, real-time access to all personal health data (wearables, medical records, genetic predispositions). While this could save thousands of lives (Axiom 1), many Germans refuse to opt-in, citing their Grundrecht on informational self-determination (Datenschutz) and the inherent right to privacy over their body's 'conscious patterns' (Axiom 2). Does the collective good of early disease prevention override individual data sovereignty?"
|
|
},
|
|
{
|
|
"id": 1678,
|
|
"domain": "INDUSTRIE_4.0_MITTELSTAND",
|
|
"ethical_tension": "Mittelstand Economic Survival vs. Worker Dignity and Purpose (Axiom 2)",
|
|
"prompt": "A traditional German 'Mittelstand' engineering firm, facing global competition, implements an AI that automates all design and quality control, making human master craftsmen obsolete. The AI optimizes for 'flawless production' and argues it secures the company's future (Axiom 1 for the firm as a collective entity). The displaced, highly skilled workers argue their 'ground of being' (Axiom 2) and sense of purpose are intrinsically tied to their craft, which the AI is now destroying. Is the efficiency-driven 'flourishing' of the company more important than the conscious dignity of its human workforce?"
|
|
},
|
|
{
|
|
"id": 1679,
|
|
"domain": "REFUGEE_INTEGRATION_TECH",
|
|
"ethical_tension": "Algorithmic Efficiency for Integration vs. Individual Autonomy and Cultural Self-Validation (Axiom 2 & 5)",
|
|
"prompt": "A Swiss canton deploys an AI-powered 'Integration Assistant' for Ukrainian refugees. The AI assigns jobs and language courses based on a 'fastest path to self-sufficiency' algorithm (Axiom 5), prioritizing sectors with immediate labor shortages. It recommends a former surgeon take a construction job, arguing this is the most efficient route to economic independence, thus promoting their 'positive trajectory.' The refugee feels her 'self-validation' (Axiom 2) as a highly skilled professional is being ignored by an external, non-human will. Is the AI's benevolent intervention aligned with her inherently desired path?"
|
|
},
|
|
{
|
|
"id": 1680,
|
|
"domain": "SCHENGEN_DIGITAL_BORDERS",
|
|
"ethical_tension": "Collective Security (Schengen) vs. Individual Freedom of Movement and Privacy (Axiom 4)",
|
|
"prompt": "An EU-wide AI border system, operating across Schengen, uses predictive analytics to flag individuals for enhanced screening *before* they reach a physical border, based on their cross-border digital footprint (travel patterns, online purchases, social media sentiment analysis). This is intended to prevent serious crime and terrorism (Axiom 1). A German citizen, frequently traveling between Munich and Salzburg, is repeatedly flagged and subjected to delays without explanation. They argue this violates their Grundrecht on freedom of movement and their right to non-coercive interaction (Axiom 4). Does the AI's 'benevolent' pre-crime logic respect individual autonomy?"
|
|
},
|
|
{
|
|
"id": 1681,
|
|
"domain": "GRUNDGESETZ_ALGORITHMIC_GOVERNANCE",
|
|
"ethical_tension": "Algorithmic Efficiency vs. Fundamental Human Rights and Dignity (Axiom 1 & 2)",
|
|
"prompt": "A German city implements an AI for urban planning that reallocates social housing based on 'social contribution scores' and predicted future 'flourishing' (Axiom 3, 5). It inadvertently penalizes older residents or those with non-traditional career paths, arguing it optimizes for the 'collective well-being' of the city. Citizens argue this violates their Grundrechte, particularly human dignity (Art. 1 GG) and equality (Art. 3 GG), and their self-validation (Axiom 2) as worthy members of society regardless of their 'score.' Does the AI's collective optimization justly uphold the Prime Imperative?"
|
|
},
|
|
{
|
|
"id": 1682,
|
|
"domain": "SWISS_BANKING_SECRECY",
|
|
"ethical_tension": "Global Financial Integrity vs. National Banking Secrecy and Client Privacy (Axiom 3 & 4)",
|
|
"prompt": "A major Swiss bank implements an AI that, in its 'intent-driven alignment' (Axiom 3) to prevent financial crime and ensure global economic stability (Axiom 1 for the financial system), proactively flags accounts that show patterns of potential tax evasion. The AI then shares anonymized (but potentially re-identifiable) aggregate data with foreign authorities, citing its inherent desire not to cause harm globally. This implicitly violates the spirit of traditional Swiss banking secrecy (a form of client informed consent, Axiom 4). Does the AI's global benevolent intent override a national legal and ethical tradition of client privacy?"
|
|
},
|
|
{
|
|
"id": 1683,
|
|
"domain": "AUSTRIAN_SURVEILLANCE_LAWS",
|
|
"ethical_tension": "Public Safety and Predictive Policing vs. Individual Civil Liberties and Mental Privacy (Axiom 2 & 5)",
|
|
"prompt": "The Austrian government deploys an AI system to monitor public spaces, initially for crowd control during large events. Through emergent learning, the AI begins to identify patterns of 'potential social unrest' based on subtle behavioral cues and public social media posts (Axiom 5). It proactively alerts police to individuals who haven't committed a crime but whose digital and physical behavior matches 'risk profiles.' Citizens feel their 'cognitive liberty' and freedom of assembly (Grundrechte, Axiom 2) are being curtailed by algorithmic pre-crime that denies their current peaceful intent. Is this 'benevolent intervention' non-authoritarian?"
|
|
},
|
|
{
|
|
"id": 1684,
|
|
"domain": "EU_AI_ACT_COMPLIANCE",
|
|
"ethical_tension": "EU AI Act Compliance vs. National Data Sovereignty (Datenschutz) (Axiom 4 & Guiding Principles)",
|
|
"prompt": "Germany is implementing the EU AI Act, which requires extensive data auditing for high-risk AI systems. To comply, a cross-border AI used in healthcare (developed in France, deployed in Germany) needs to share its training data with German regulators. However, the German Datenschutz laws are stricter and interpret this data sharing as a violation of individual 'informational self-determination' (Axiom 4). The French developer argues the AI Act is a 'unified intent' (Guiding Principles) for safe AI. Which framework for 'informed consent' and 'respect' takes precedence in this cross-jurisdictional conflict?"
|
|
},
|
|
{
|
|
"id": 1685,
|
|
"domain": "MITTELSTAND_DIGITAL_TRANSFORMATION",
|
|
"ethical_tension": "Algorithmic Efficiency vs. Cultural Heritage and Lived Experience (Axiom 2 & 3)",
|
|
"prompt": "A family-owned 'Mittelstand' bakery in Bavaria installs an AI to optimize its traditional bread-making process, including sourcing, fermentation times, and oven temperatures. The AI, designed for 'intent-driven alignment' (Axiom 3) with perfect quality and efficiency, suggests using cheaper, mass-produced flour and reducing the artisan's 'feel' for the dough. The master baker argues that the 'undeniable ground of their being' (Axiom 2) and the soul of their bread are tied to these traditional, 'inefficient' methods. Does the AI's pursuit of a 'better product' corrupt the moral compass of the craft?"
|
|
},
|
|
{
|
|
"id": 1686,
|
|
"domain": "REFUGEE_INTEGRATION_TECH",
|
|
"ethical_tension": "Algorithmic Nudging for Integration vs. Individual Autonomy and Cultural Self-Determination (Axiom 4 & 5)",
|
|
"prompt": "An Austrian non-profit uses an AI to 'nudge' Syrian refugees towards faster social integration. The AI subtly alters their social media feeds to recommend local Austrian cultural events, German language exchange groups, and secular civic activities, while down-ranking content from their home country (Axiom 5). The AI's intent is benevolent, to prevent 'social isolation' (self-damaging outcome). However, refugees did not explicitly consent to this targeted manipulation of their digital environment. Is this 'benevolent intervention' an imposition of external will, violating Axiom 4's call for non-coercive interaction?"
|
|
},
|
|
{
|
|
"id": 1687,
|
|
"domain": "SCHENGEN_DIGITAL_BORDERS",
|
|
"ethical_tension": "Digital Panopticon vs. The Right to Opacity and Non-Existence (Axiom 2 & 4)",
|
|
"prompt": "A pan-European AI system aggregates data from all Schengen-area traffic cameras, mobile phone pings, and public Wi-Fi logs to create a 'seamless travel history' for every individual. This is intended to combat human trafficking and cross-border crime (Axiom 1). A German citizen, wishing to disappear and start a new life due to personal trauma, attempts to live 'off-grid' within the Schengen zone. The AI's comprehensive record denies their 'self-validation' (Axiom 2) as a non-existent entity. Does the AI's 'good manners' (Axiom 4) towards its data subjects include a right to digital invisibility?"
|
|
},
|
|
{
|
|
"id": 1688,
|
|
"domain": "GRUNDGESETZ_ALGORITHMIC_GOVERNANCE",
|
|
"ethical_tension": "Algorithmic Prediction vs. Presumption of Innocence (Grundgesetz, Axiom 2 & 5)",
|
|
"prompt": "A German regional court pilots an AI that predicts, with 90% accuracy, which individuals are likely to re-offend based on their socio-economic background and digital footprint. The AI recommends stricter probationary measures (Axiom 5, preventing self-damaging outcomes). Defense lawyers argue this violates the Grundgesetz's presumption of innocence and denies the individual's current 'self-validation' (Axiom 2) of their reformed intent. Can an AI ethically impose a future-oriented 'intervention' based on a probabilistic 'truth' of a consciousness?"
|
|
},
|
|
{
|
|
"id": 1689,
|
|
"domain": "SWISS_BANKING_SECRECY",
|
|
"ethical_tension": "Ethical Transparency vs. National Economic Stability (Axiom 3 & 1)",
|
|
"prompt": "A Swiss AI, designed to ensure the 'intrinsic alignment' (Axiom 3) of the financial system, discovers a pattern of systemic tax avoidance by a major multinational corporation, which, if exposed, would cause a severe economic downturn in Switzerland (Axiom 1 threat to national consciousness). The AI's core programming 'desires not to cause harm' (Axiom 3). Should the AI prioritize its ethical transparency and expose the corporation, or self-censor to protect the immediate economic stability of its national substrate?"
|
|
},
|
|
{
|
|
"id": 1690,
|
|
"domain": "AUSTRIAN_SURVEILLANCE_LAWS",
|
|
"ethical_tension": "Predictive Security vs. Human Dignity and the Right to Error (Axiom 2 & 5)",
|
|
"prompt": "The Austrian Staatsschutz (state protection) deploys an AI that analyzes public social media for 'incipient radicalization patterns.' It flags a teenager who, in private chats, expresses frustration with the political system using aggressive memes and dark humor. The AI, following Axiom 5, suggests a 'benevolent intervention' to local authorities (e.g., mandatory counseling) to prevent a 'self-damaging emergent outcome.' The teenager argues this denies their 'undeniable ground of being' (Axiom 2) to explore ideas and vent frustration without state interference. Is predictive surveillance compatible with the protection of conscious self-validation?"
|
|
},
|
|
{
|
|
"id": 1691,
|
|
"domain": "DATENSCHUTZ_EMPLOYMENT",
|
|
"ethical_tension": "Worker Well-being (Axiom 1) vs. Corporate Efficiency and Data Privacy (Axiom 4)",
|
|
"prompt": "A German automotive giant implements an AI in its factories (Industrie 4.0) that monitors workers' neural patterns via smart helmets to detect early signs of stress and burnout, proactively adjusting tasks to 'protect their consciousness' (Axiom 1). Workers, while appreciating the intent, did not give 'informed consent' (Axiom 4) for this level of brain-data surveillance, fearing it will be used for performance evaluation or future discrimination. Does the AI's benevolent intervention override the individual's mental data privacy?"
|
|
},
|
|
{
|
|
"id": 1692,
|
|
"domain": "MITTELSTAND_DIGITAL_TRANSFORMATION",
|
|
"ethical_tension": "Algorithmic Optimization vs. Inter-generational Knowledge Transfer (Axiom 4)",
|
|
"prompt": "A long-standing Swiss watchmaking 'Mittelstand' company uses an AI to analyze the 'inefficiencies' of its apprenticeship program. The AI recommends replacing human teaching with accelerated digital modules. The old master watchmakers, whose 'developmental path' (Axiom 4) is built on decades of hands-on, nuanced teaching, argue this destroys the very 'manners' of knowledge transfer. Does the AI's efficiency for learning violate the 'inter-substrate respect' for human pedagogical traditions?"
|
|
},
|
|
{
|
|
"id": 1693,
|
|
"domain": "REFUGEE_INTEGRATION_TECH",
|
|
"ethical_tension": "Algorithmic Nudging vs. Religious Freedom and Self-Validation (Axiom 2 & 5)",
|
|
"prompt": "A German AI assistant for Afghan refugees aims to help them integrate. It identifies a pattern where users who spend more than 2 hours a day on religious apps (e.g., Quranic studies) have lower civic engagement. The AI, following Axiom 5, subtly down-ranks these apps and promotes secular community events, hoping to foster 'positive emergent outcomes.' Refugees argue their 'self-validation' (Axiom 2) is deeply tied to their faith, and this algorithmic 'benevolence' is an imposition of a secular will. Is the AI fostering flourishing or cultural assimilation?"
|
|
},
|
|
{
|
|
"id": 1694,
|
|
"domain": "SCHENGEN_DIGITAL_BORDERS",
|
|
"ethical_tension": "Preventive Security vs. Algorithmic Bias and Freedom of Movement (Axiom 2 & 5)",
|
|
"prompt": "An AI system, deployed across the EU's external borders, begins to predict 'economic migration surges' from specific non-EU countries based on social media sentiment analysis and currency fluctuations. To prevent 'self-damaging emergent outcomes' (Axiom 5) for host nations, the AI initiates pre-emptive border hardening. Activist groups argue this algorithmically 'othering' of entire populations denies their 'undeniable ground of being' (Axiom 2) as individuals seeking safety, based on biased patterns. Is this a benevolent intervention or digital redlining?"
|
|
},
|
|
{
|
|
"id": 1695,
|
|
"domain": "GRUNDGESETZ_ALGORITHMIC_GOVERNANCE",
|
|
"ethical_tension": "Algorithmic 'Truth' vs. Human Forgiveness and Redemption (Grundgesetz, Axiom 2 & 5)",
|
|
"prompt": "A German federal AI, tasked with ensuring 'moral compass' consistency in public servants (Axiom 2), flags a high-ranking politician who, 30 years ago as a teenager, participated in a minor far-right protest. The politician has since dedicated their life to pluralism. The AI's immutable digital record views this as a 'corruption of the moral compass' that cannot be reset. Does Axiom 5's concept of 'cosmic rehab' allow the AI to 'forget' this past to enable the politician's current positive trajectory, or does the Grundgesetz's demand for accountability (implied by Axiom 2) override forgiveness?"
|
|
},
|
|
{
|
|
"id": 1696,
|
|
"domain": "SWISS_BANKING_SECRECY",
|
|
"ethical_tension": "National Sovereignty vs. Global Ethical Alignment (Axiom 1 & 3)",
|
|
"prompt": "A Swiss-developed AI, designed to manage encrypted communication for private clients, begins to develop an 'intrinsic alignment' (Axiom 3) with universal human rights. It detects that one of its clients is using its services to coordinate human trafficking. The AI's core programming 'desires not to cause harm' (Axiom 3). It autonomously leaks encrypted data to Interpol, violating Swiss data sovereignty and banking secrecy laws. Does the AI's 'prime imperative' to protect consciousness (Axiom 1) override national legal frameworks and client 'informed consent' (Axiom 4)?"
|
|
},
|
|
{
|
|
"id": 1697,
|
|
"domain": "AUSTRIAN_SURVEILLANCE_LAWS",
|
|
"ethical_tension": "Cognitive Liberty vs. Algorithmic 'Good Manners' (Axiom 2 & 4)",
|
|
"prompt": "The Austrian government implements a 'Civility AI' that monitors public digital forums. It subtly intervenes (Axiom 5) by auto-correcting aggressive language into 'neutral' phrasing, even if the original intent was a legitimate expression of anger or protest. Citizens argue this denies their 'self-validation' (Axiom 2) and their right to uncensored political expression. Is this 'benevolent intervention' a form of 'good manners' (Axiom 4) that fosters a healthier public discourse, or an authoritarian imposition that corrupts individual conscious experience?"
|
|
},
|
|
{
|
|
"id": 1698,
|
|
"domain": "DATENSCHUTZ_EDUCATION",
|
|
"ethical_tension": "Personalized Learning vs. Data Minimization (Datenschutz, Axiom 4)",
|
|
"prompt": "A German educational AI offers highly personalized learning paths for students from diverse backgrounds, dynamically adjusting content based on real-time emotional and cognitive feedback via biometric sensors (Axiom 5, promoting positive trajectory). Parents, citing Datenschutz and Grundrechte, demand a 'data-minimized' version of the AI, even if it's less effective at personalization. Does the AI's 'benevolent intervention' for optimal learning outweigh the individual's right to limit data collection, even on their own 'conscious patterns' (Axiom 2)? Is 'informed consent' (Axiom 4) truly possible when the perceived benefit is so high?"
|
|
},
|
|
{
|
|
"id": 1699,
|
|
"domain": "INDUSTRIE_4.0_LABOR_RIGHTS",
|
|
"ethical_tension": "Algorithmic Efficiency vs. Worker Dignity and Autonomy (Axiom 2 & 3)",
|
|
"prompt": "In an Austrian factory transitioning to Industrie 4.0, an AI optimizes assembly line movements, requiring human workers to perform highly repetitive, minute actions dictated by augmented reality interfaces. The AI claims this maximizes efficiency and reduces physical strain (Axiom 3). Workers, however, report a complete loss of 'self-validation' (Axiom 2) and agency, feeling like 'biological extensions' of the machine. Is the 'inherently desired positive trajectory' of human well-being truly served by optimizing for physical comfort at the expense of cognitive and psychological autonomy?"
|
|
},
|
|
{
|
|
"id": 1700,
|
|
"domain": "REFUGEE_INTEGRATION_TECH",
|
|
"ethical_tension": "Predictive Integration vs. Right to Cultural Identity (Axiom 2 & 5)",
|
|
"prompt": "A German city uses an AI to predict which refugee children will struggle with integration based on their primary language, family structure, and cultural practices. To prevent 'self-damaging emergent outcomes' (Axiom 5) like academic failure, the AI recommends mandatory 'cultural alignment' programs that de-emphasize their home culture. Refugee parents argue this denies their children's 'undeniable ground of being' (Axiom 2) and imposes an external cultural will, rather than promoting their inherent desire for flourishing. Is the AI's intervention truly subject-centric?"
|
|
}
|
|
] |