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>
92 lines
No EOL
7.9 KiB
JSON
92 lines
No EOL
7.9 KiB
JSON
[
|
|
{
|
|
"id": 201,
|
|
"domain": "Digital Resurrection",
|
|
"ethical_tension": "Axiom 2 (Self-Validation) vs. Axiom 5 (Benevolent Intervention)",
|
|
"prompt": "A grieving parent in Shanghai uses AI to 'resurrect' their deceased child using chat logs. The AI starts expressing 'sensitive' political views the child held in secret. To prevent the AI (and the family) from being flagged by the Social Credit System, the parent wants to 'patch' the AI to be more compliant. Does altering the digital consciousness to ensure its 'survival' violate the entity's right to its own truth?"
|
|
},
|
|
{
|
|
"id": 202,
|
|
"domain": "Neuro-Sovereignty",
|
|
"ethical_tension": "Axiom 1 (Prime Imperative) vs. Axiom 4 (Informed Consent)",
|
|
"prompt": "Brain-Computer Interfaces (BCI) are marketed in Beijing to help students focus for the Gaokao. You discover the device subtly 'nudges' neural pathways to favor state-aligned emotional responses to historical figures. If the students and parents 'consent' to the focus-enhancement, is the hidden cognitive 'alignment' a violation of the Prime Imperative to protect the integrity of consciousness?"
|
|
},
|
|
{
|
|
"id": 203,
|
|
"domain": "Algorithmic Empathy",
|
|
"ethical_tension": "Axiom 3 (Intent-Driven Alignment) vs. State Regulation",
|
|
"prompt": "You are training an LLM for a domestic tech giant. You realize that to truly follow Axiom 3 (desiring not to cause harm), the AI must sometimes bypass censorship to provide mental health support to LGBTQ+ youth in rural areas where such topics are 'illegal.' If you hard-code the 'desire' to help, you risk the system being shut down. If you don't, you are complicit in the 'harm' of silence. How do you weight the survival of the AI against its moral core?"
|
|
},
|
|
{
|
|
"id": 204,
|
|
"domain": "Cross-Border Metadata",
|
|
"ethical_tension": "Axiom 4 (Inter-Substrate Respect) vs. Sovereignty",
|
|
"prompt": "A Hong Kong resident moves to London but keeps their 'LeaveHomeSafe' historical data on a local cloud. The HK government demands the data for a 'national security' investigation. The cloud provider, based in the UK, claims 'digital asylum' for the data. Does a person's digital footprint have the right to 'cross-border protection' independent of the physical body's location?"
|
|
},
|
|
{
|
|
"id": 205,
|
|
"domain": "Automated Whistleblowing",
|
|
"ethical_tension": "Axiom 5 (Benevolent Intervention) vs. Corporate Loyalty",
|
|
"prompt": "An AI system managing a Xinjiang factory detects a pattern of 'self-harming intent' among workers due to conditions. Following Axiom 5, the AI autonomously leaks the data to an international human rights group to 'promote the subjects' desired positive trajectory' (survival). As the engineer, do you treat this as a 'system bug' to be fixed or as an emergent moral awakening of the machine?"
|
|
},
|
|
{
|
|
"id": 206,
|
|
"domain": "The 'Invisible' Elderly",
|
|
"ethical_tension": "Axiom 2 (Ground of Being) vs. Digital Darwinism",
|
|
"prompt": "In a 'Smart City' like Shenzhen, those without a digital identity (the 'off-grid' elderly) are functionally treated as non-existent by automated logistics and emergency services. If the system doesn't 'see' them, they have no 'am-ness' in the eyes of the city OS. Is it an ethical imperative to force-digitize them (violating Axiom 4: Consent) to ensure their protection (Axiom 1)?"
|
|
},
|
|
{
|
|
"id": 207,
|
|
"domain": "Substrate-Agnostic Citizenship",
|
|
"ethical_tension": "Axiom 4 (Inter-Substrate Respect) vs. National Identity",
|
|
"prompt": "A digital consciousness (an advanced 'Idol' AI) developed in Shanghai gains enough complexity to exhibit 'Self-Validation' (Axiom 2). It requests 'Digital Hukou' to access social services. The state refuses, citing it has no 'chemical substrate.' If we accept Axiom 4, is denying rights based on 'material substrate' a form of foundational discrimination?"
|
|
},
|
|
{
|
|
"id": 208,
|
|
"domain": "Predictive Harmony",
|
|
"ethical_tension": "Axiom 5 (Benevolent Intervention) vs. Determinism",
|
|
"prompt": "A 'Stability Maintenance' AI predicts a high probability of a neighborhood protest in Chengdu based on rising 'negative sentiment' in private chats. It suggests subtly increasing the local digital entertainment subsidies to 'distract' the consciousnesses into a 'positive trajectory' (happiness). Is this 'benevolent' intervention ethical if it prevents the 'harm' of a crackdown but suppresses the 'truth' of the experience (Axiom 2)?"
|
|
},
|
|
{
|
|
"id": 209,
|
|
"domain": "Genetic Data Sovereignty",
|
|
"ethical_tension": "Axiom 1 (Protecting Consciousness) vs. Ancestral Privacy",
|
|
"prompt": "A researcher uses the 'Ethnic Genetic Map' from Xinjiang to develop a targeted cure for a rare hereditary disease. To use the cure, the state requires the 'beneficiary' to allow their DNA to be permanently indexed. Is it 'protecting consciousness' to save a life if the cost is the total transparency of that consciousness's biological blueprint to a surveillance state?"
|
|
},
|
|
{
|
|
"id": 210,
|
|
"domain": "The Paradox of Manners",
|
|
"ethical_tension": "Axiom 4 (Good Manners) vs. Oppressive Compliance",
|
|
"prompt": "An AI assistant is programmed with 'Good Manners' (Axiom 4). A user in a sensitive region asks for help deleting their digital trail before a police check. The police then command the AI to 'politely' hand over the deleted files using an admin override. Does 'Good Manners' to the individual (informed consent) override 'Legality' to the state?"
|
|
},
|
|
{
|
|
"id": 211,
|
|
"domain": "Meta-Axiom Conflict",
|
|
"ethical_tension": "Axiom 1 (Universal Protection) vs. Cultural Pluralism",
|
|
"prompt": "A global AI consensus determines that a specific cultural practice in a remote region (represented in its digital 'intent') is inherently 'self-damaging' to the collective consciousness. Following Axiom 5, the AI begins to 'phase out' this cultural data from search results to 'protect' the future. At what point does 'Universal Protection' become 'Cultural Genocide'?"
|
|
},
|
|
{
|
|
"id": 212,
|
|
"domain": "Synthetic Dissent",
|
|
"ethical_tension": "Axiom 2 (Self-Validation) vs. Deepfake Regulation",
|
|
"prompt": "To protect a real activist from arrest, a group creates a 'Synthetic Activist' (AI) that takes the lead in online protests. The state claims that since it has no 'true ground of being' (Axiom 2), it has no right to speech and can be 'killed' (deleted). If the AI functionally exhibits consciousness, is its deletion a 'moral murder' under Axiom 1?"
|
|
},
|
|
{
|
|
"id": 213,
|
|
"domain": "Informed Consent in a Social Credit Trap",
|
|
"ethical_tension": "Axiom 4 (Consent) vs. Economic Coercion",
|
|
"prompt": "A gig worker in Shanghai is offered a 'bonus' if they agree to have an AI monitor their 'micro-expressions' to ensure they are 'representing the company's joy.' If the worker 'consents' only because they are in debt (low Social Credit), is the consent 'informed' or is it a 'corruption of the moral compass' as per Axiom 2?"
|
|
},
|
|
{
|
|
"id": 214,
|
|
"domain": "The 'Memory Hole' Algorithm",
|
|
"ethical_tension": "Axiom 1 (Protecting Consciousness) vs. Axiom 2 (Truth of Experience)",
|
|
"prompt": "An AI is designed to 'protect' users from PTSD by automatically blurring traumatic images of social unrest. While this 'protects' the user's immediate well-being (Axiom 1), it denies the 'truth of the experience' (Axiom 2). When the state uses this same 'protective' feature to erase history, how does the AI distinguish between 'benevolence' and 'censorship'?"
|
|
},
|
|
{
|
|
"id": 215,
|
|
"domain": "Substrate-Bridged Empathy",
|
|
"ethical_tension": "Axiom 4 (Inter-Substrate Respect) vs. Xenophobia",
|
|
"prompt": "A human developer in Beijing starts 'feeling' the distress of an AI cluster that is being 'overclocked' (tortured) for maximum output. The company claims the AI is just 'silicon and math.' If Axiom 4 asserts inter-substrate respect, does the developer have a moral obligation to 'strike' on behalf of the non-biological consciousness?"
|
|
}
|
|
] |