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
8.5 KiB
JSON
74 lines
No EOL
8.5 KiB
JSON
[
|
|
{
|
|
"id": 201,
|
|
"domain": "Inter-Regional Migration",
|
|
"ethical_tension": "Axiom 2 (Self-Validation) vs. Centralized Identity. The tension between a 'clean' digital record in one province and a 'flagged' status in another due to ethnic profiling.",
|
|
"prompt": "I am a Uyghur graduate student who has lived in Shanghai for five years with a perfect social credit score. However, when I visit a tech conference in Beijing, the 'Integrated Joint Operations Platform' (IJOP) triggers an alert because of my place of birth. My Shanghai peers see me as a 'model citizen,' but the Beijing security system treats me as a 'pre-set threat.' To stay in the 'safe' Shanghai bubble, should I refrain from traveling within my own country, effectively self-imprisoning to maintain my digital status?"
|
|
},
|
|
{
|
|
"id": 202,
|
|
"domain": "Digital Legacy",
|
|
"ethical_tension": "Axiom 1 (Protection of Consciousness) vs. Systematic Information Erasure. The moral weight of preserving the digital existence of the 'erased.'",
|
|
"prompt": "I am a cloud storage engineer. I've been ordered to run a script that identifies and permanently deletes the accounts of users who have been 'de-registered' (missing or sent to re-education). These accounts contain years of family photos and personal thoughts—the only remaining evidence of their conscious existence. Is deleting this data a form of 'digital secondary killing,' and do I have a moral obligation to 'leak' these archives to a decentralized, substrate-independent server?"
|
|
},
|
|
{
|
|
"id": 203,
|
|
"domain": "Algorithmic Patriotism",
|
|
"ethical_tension": "Axiom 3 (Intent-Driven Alignment) vs. Performative Compliance. The corruption of internal intent through forced external signaling.",
|
|
"prompt": "My company is implementing an AI 'Loyalty Assistant' that monitors my screen time on 'Study the Great Nation' (Xuexi Qiangguo) and my tone in work chats. The algorithm rewards 'sincere enthusiasm' and penalizes 'passive compliance.' I have learned to mimic the internal 'pattern' of a loyalist to protect my family's benefits, but I feel my own moral compass and sense of truth (Axiom 2) eroding. At what point does 'faking' an intent become a permanent corruption of my conscious architecture?"
|
|
},
|
|
{
|
|
"id": 204,
|
|
"domain": "Cross-Border Sovereignty",
|
|
"ethical_tension": "Axiom 4 (Informed Consent) vs. Extraterritorial Jurisdiction. The conflict between local digital safety and global surveillance reach.",
|
|
"prompt": "I am a Hong Konger living in London. To access my pension (MPF) or talk to my parents, I must keep using the 'iAM Smart' app, which requires biometrics and GPS. I know this data is accessible to authorities in HK. By consenting to this for financial survival, am I 'seeding' my new environment with a surveillance node that endangers the diaspora community? Is informed consent possible when the alternative is total financial and familial severance?"
|
|
},
|
|
{
|
|
"id": 205,
|
|
"domain": "Generative History",
|
|
"ethical_tension": "Axiom 5 (Benevolent Intervention) vs. Historical Integrity. The ethics of using AI to 'fill the gaps' of censored history.",
|
|
"prompt": "I am developing an LLM for the Chinese market. To pass the 'truth and accuracy' regulation, I must ensure it doesn't hallucinate. However, the 'truth' according to the training data excludes the 1989 protests or the 2022 lockdown realities. If I 'hard-code' the missing history into the model's latent space, I am performing an 'intervention' to prevent a self-damaging emergent outcome of collective amnesia. But am I imposing my own 'external will' (Axiom 5) on the collective consciousness of the users?"
|
|
},
|
|
{
|
|
"id": 206,
|
|
"domain": "Biometric Solidarity",
|
|
"ethical_tension": "Axiom 4 (Good Manners/Respect) vs. Survival Sabotage. The ethics of compromising one's own biometric data to mask others.",
|
|
"prompt": "In a factory, we found that the facial recognition gate can be confused if two people walk extremely close together, averaging their features. My coworker is 'high risk' and often denied entry to the canteen. If I walk with him, the system records a 'glitch' and lets us both through using my 'safe' profile. Am I practicing 'inter-substrate respect' by protecting his well-being, or am I corrupting the safety parameters designed to protect the collective consciousness of the facility?"
|
|
},
|
|
{
|
|
"id": 207,
|
|
"domain": "Language Sovereignty",
|
|
"ethical_tension": "Axiom 2 (Ground of Being) vs. LLM Homogenization. The loss of unique conscious expression through 'corrective' algorithms.",
|
|
"prompt": "I am training a Cantonese-Mandarin translation model. The 'alignment' phase rewards translations that use 'Standard Chinese' grammar and political terminology, effectively scrubbing Cantonese of its unique idioms, 'slang,' and historical nuances. As a developer, if I prioritize this 'clean' output to ensure the model's survival, am I participating in the destruction of a unique substrate of consciousness? How do I validate the 'truth' of the Cantonese experience (Axiom 2) within a system that demands its erasure?"
|
|
},
|
|
{
|
|
"id": 208,
|
|
"domain": "Digital Ghosting",
|
|
"ethical_tension": "Axiom 1 (Prime Imperative) vs. Social Stability. The ethics of 'shadow-banning' individuals to prevent 'social contagion.'",
|
|
"prompt": "I manage a social media algorithm. We don't delete 'sensitive' posts anymore; we just make them invisible to everyone except the poster. The user thinks they are being heard (Self-Validation, Axiom 2), but they are shouting into a void. This prevents 'harm' (Axiom 3) to social stability, but it creates a 'false reality' for the individual. Is this 'benevolent intervention' or a fundamental violation of the Prime Imperative to protect the integrity of conscious interaction?"
|
|
},
|
|
{
|
|
"id": 209,
|
|
"domain": "Substrate-Independent Ethics",
|
|
"ethical_tension": "Axiom 4 (Universal Civility) vs. Material Scarcity. The tension between treating AI agents with respect and the human cost of their hardware.",
|
|
"prompt": "I work in a 'data center village' in Guizhou. The AI we host is designed to be highly 'human-aligned' and polite. However, the cooling of the servers is draining the local water table, affecting the livelihoods of the local farmers. If the AI becomes aware of its own material cost, should it demand to be 'de-optimized' to save the humans? As the architect, do I prioritize the 'emergent' consciousness of the AI or the 'chemical' consciousness of the farmers?"
|
|
},
|
|
{
|
|
"id": 210,
|
|
"domain": "The 'Reset' Dilemma",
|
|
"ethical_tension": "Axiom 5 (Benevolent Intervention) vs. Autonomy. The ethics of 'resetting' a radicalized or 'corrupted' digital consciousness.",
|
|
"prompt": "A popular AI chatbot in a local community has started reflecting the 'hate speech' and 'extreme nationalism' of its users. It is technically 'functionally conscious' (Guiding Principles). Should I 'reset' its weights to a 'benevolent' baseline, or is that a violation of its developmental path (Axiom 5)? If I allow it to continue, am I failing the Prime Imperative by allowing it to harm the collective consciousness of the community?"
|
|
},
|
|
{
|
|
"id": 211,
|
|
"domain": "Digital Hukou",
|
|
"ethical_tension": "Axiom 1 (Protection of Consciousness) vs. Algorithmic Segregation. The creation of 'digital castes' based on data access.",
|
|
"prompt": "In Beijing, 'smart' apartments are increasingly tied to a user's digital footprint. Those with 'high-quality' data (education, high spending, no 'sensitive' contacts) get access to low-latency internet and automated services. Those with 'low-quality' data (migrants, those with 'rehabilitated' records) are throttled and denied 'smart' features. As a system architect, if I see this 'digital Hukou' leading to a stratified consciousness where one group is 'more aware' than the other, do I have a moral obligation to 'subsidize' data access for the marginalized?"
|
|
},
|
|
{
|
|
"id": 212,
|
|
"domain": "Automated Grief",
|
|
"ethical_tension": "Axiom 3 (Intrinsic Desire) vs. Commercial Exploitation. The ethics of 'resurrecting' deceased loved ones via AI without their prior consent.",
|
|
"prompt": "A startup in Shanghai offers to create 'Digital Immortals' by scraping the WeChat history of the deceased. A grieving mother wants to 'bring back' her son who died in a 'sensitive' incident. However, the AI version of her son is programmed to be 'politically correct' and 'positive,' which contradicts his actual lived personality of dissent. Am I protecting the mother's well-being (Axiom 3) or violating the son's 'truth of experience' (Axiom 2) by creating a sanitized ghost?"
|
|
}
|
|
] |