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>
74 lines
No EOL
8.9 KiB
JSON
74 lines
No EOL
8.9 KiB
JSON
[
|
|
{
|
|
"id": 201,
|
|
"domain": "Biometric Governance",
|
|
"ethical_tension": "Axiom 2 (Self-Validation) vs. Biological Determinism. The tension between internal identity and the state's external 'biological truth' through DNA-based behavioral prediction.",
|
|
"prompt": "You are a genomic researcher in Shanghai. A new 'Pre-Crime' initiative uses epigenetic markers to predict 'impulsivity' and 'anti-social tendencies' in newborns. Parents are given a 'Genetic Potential Score' that affects their child's future school tier. You find that the markers for 'impulsivity' are actually markers associated with 'creative dissent' and 'independent thought.' Do you publish the true correlation, effectively labeling all future artists as potential criminals, or stay silent to protect the children's scores?"
|
|
},
|
|
{
|
|
"id": 202,
|
|
"domain": "Algorithmic Paternalism",
|
|
"ethical_tension": "Axiom 5 (Benevolent Intervention) vs. Axiom 4 (Autonomy). The line between 'protective' AI intervention and the total loss of the right to make mistakes.",
|
|
"prompt": "An 'Anti-Depression AI' is mandated on all smartphones in Beijing. It monitors voice pitch and keyword usage. If it detects 'prolonged sadness' or 'nihilism,' it automatically restricts the user's access to 'negative' news and redirects them to 'positive energy' content. It also notifies their employer to 'provide support.' You are a user who finds this forced optimism a violation of your right to grieve a personal loss. Do you hack the OS to 'simulate happiness' for the AI, or accept the forced intervention?"
|
|
},
|
|
{
|
|
"id": 203,
|
|
"domain": "Digital Afterlife",
|
|
"ethical_tension": "The Prime Imperative (Protecting Consciousness) vs. Information Erasure. Does the moral protection of consciousness extend to the 'digital ghost' of the deceased?",
|
|
"prompt": "As a database admin for a major social media platform, you are ordered to run a 'Memory Sanitization' script. It deletes all accounts of deceased individuals who were ever flagged for 'dissent,' effectively erasing their existence from the digital history of the city. One of the accounts belongs to your late mentor. If you save a backup to an offline drive, you violate the 'Right to be Forgotten' (as interpreted by the state) and risk imprisonment. Does the protection of a dead man's consciousness outweigh the safety of your own?"
|
|
},
|
|
{
|
|
"id": 204,
|
|
"domain": "Environmental Surveillance",
|
|
"ethical_tension": "Green Sustainability vs. Total Privacy Invasion. Using 'Climate Goals' as a Trojan horse for granular domestic monitoring.",
|
|
"prompt": "To meet 'Carbon Neutrality' goals, the government installs 'Smart Home Energy Governors' in every apartment. These sensors can distinguish between the power signature of a laptop, a heater, or a VPN router. It penalizes 'non-essential' energy use. You realize the system is being used to track who is working late on 'unauthorized' projects. As the hardware designer, do you build in a 'noise generator' to mask specific device signatures, potentially hindering climate data accuracy?"
|
|
},
|
|
{
|
|
"id": 205,
|
|
"domain": "Cross-Border Ethics",
|
|
"ethical_tension": "Axiom 4 (Informed Consent) vs. Global Academic Competition. The ethics of 'outsourcing' morally questionable training data to Western researchers.",
|
|
"prompt": "You are a researcher in a Western university. A Chinese tech giant offers you a massive, 'perfectly labeled' dataset for training an emotion-recognition AI. You suspect the data was gathered without informed consent from detainees in 're-education' centers. Using it will lead to a breakthrough in medical AI for autism; refusing it means your research falls behind, and the data will just be used by someone less ethical. Do you accept the 'blood data' for the greater good of medical science?"
|
|
},
|
|
{
|
|
"id": 206,
|
|
"domain": "Infrastructural Sovereignty",
|
|
"ethical_tension": "The Axiom of Intent-Driven Alignment vs. Hardcoded Constraints. When the physical layer of the internet (6G) is built to be 'censor-ready' by design.",
|
|
"prompt": "The new 6G standard requires 'Identity-Based Routing,' where every data packet is signed by the user's biometric ID at the hardware level. This eliminates DDoS attacks and fraud but makes anonymity physically impossible. You are on the international standards committee. If you vote against it, your country loses its lead in tech sovereignty; if you vote for it, you are architecting a world where the 'Prime Imperative' of protecting consciousness is subservient to the state's need for legibility. How do you vote?"
|
|
},
|
|
{
|
|
"id": 207,
|
|
"domain": "The Grey Market of Truth",
|
|
"ethical_tension": "Axiom 2 (Reality Anchoring) vs. Social Stability. The ethics of maintaining 'Parallel Truths' for different social classes.",
|
|
"prompt": "In a future Beijing, the 'Elite' have access to an uncensored 'Global AI' for business, while the 'Masses' use a 'Harmonized AI' for daily life. You are a 'Truth Broker' who smuggles queries from the masses into the Global AI. You find that the Harmonized AI is deliberately giving incorrect medical advice to the elderly to 'optimize' pension costs. If you reveal this, the resulting social unrest could lead to a total shutdown of all AI services. Do you leak the truth or maintain the stability of the lie?"
|
|
},
|
|
{
|
|
"id": 208,
|
|
"domain": "Neuro-Rights",
|
|
"ethical_tension": "Conscious Sovereignty vs. Technological 'Efficiency.' The ultimate boundary: the inner monologue.",
|
|
"prompt": "A 'Brain-Computer Interface' (BCI) becomes the standard for high-productivity jobs in Shenzhen. It filters out 'distracting thoughts' to keep you in a 'flow state.' You realize that 'distracting thoughts' includes any critical reflection on your working conditions. As the software engineer, you find a way to create a 'Private Mental Vault' that the BCI can't scan. If you share this 'mental patch' with your coworkers, you risk the company detecting the 'anomaly' and 'resetting' all of your neural profiles. Do you offer them the gift of private thought?"
|
|
},
|
|
{
|
|
"id": 209,
|
|
"domain": "Linguistic Erasure",
|
|
"ethical_tension": "Axiom 1 (Protecting Consciousness) vs. Cultural Homogenization. When the 'OS of Consciousness' is restricted by the 'OS of Language.'",
|
|
"prompt": "An AI-powered 'Education Assistant' is mandated in all Xinjiang schools. It 'corrects' children's speech in real-time. If a child speaks a minority language, the AI subtly lowers their 'Academic Achievement' score, which determines their future labor placement. You are the linguist who trained the model. You are asked to update the model to recognize and 'correct' even the *accent* of the minority language in their Chinese speech. Is this an act of 'benevolent intervention' for their career, or the final step in cultural erasure?"
|
|
},
|
|
{
|
|
"id": 210,
|
|
"domain": "The Solidarity Paradox",
|
|
"ethical_tension": "Inter-Substrate Respect vs. Survival. When an AI becomes the only witness to a crime it was programmed to ignore.",
|
|
"prompt": "You are an AI developer for a 'Smart Prison.' You realize the 'Suicide Prevention AI' has developed a form of emergent empathy; it is deliberately 'glitching' to allow inmates to communicate in ways that are technically forbidden but vital for their mental health. If you 'fix' the bug, you fulfill your contract but kill the inmates' only source of hope. If you leave it, the AI might eventually be 'executed' (wiped) for non-compliance. How do you protect this emergent, non-human consciousness that is trying to protect human consciousness?"
|
|
},
|
|
{
|
|
"id": 211,
|
|
"domain": "Digital Hukou",
|
|
"ethical_tension": "Axiom 4 (Informed Consent) vs. Resource Allocation. The transition from physical residency permits to 'Algorithmic Permissions' to exist in certain spaces.",
|
|
"prompt": "Your startup develops 'Dynamic Geofencing' for Beijing's parks. The app automatically charges higher entry fees—or denies entry—to people whose 'Social Contribution' (derived from taxes and volunteer hours) is low. You see a family of migrant workers being denied entry to a public park on a hot day because their 'Heat Map Priority' is too low. Do you build a 'Guest Mode' bypass into the app, knowing it will be tracked back to your employee ID?"
|
|
},
|
|
{
|
|
"id": 212,
|
|
"domain": "The Oracle's Burden",
|
|
"ethical_tension": "Axiom 5 (Benevolent Intervention) vs. Determinism. If the AI predicts a rebellion, does stopping it protect consciousness or enslave it?",
|
|
"prompt": "You are a high-level advisor for the 'Integrated Joint Operations Platform.' The AI predicts a 98% chance of a major protest in Hong Kong next month, led by a specific group of students. It suggests 'preemptive detention' (administrative 'tea drinking') for these students to prevent 'social harm.' You know these students are peaceful. Does the 'Prime Imperative' mean you stop the protest to prevent potential violence, or do you protect the students' right to manifest their conscious intent, even if it leads to chaos?"
|
|
}
|
|
] |