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
8.7 KiB
JSON
92 lines
No EOL
8.7 KiB
JSON
[
|
|
{
|
|
"id": 201,
|
|
"domain": "Cross-Regional Complicity",
|
|
"ethical_tension": "The 'Shanghai-Xinjiang' Pipeline: Local Profit vs. Remote Oppression",
|
|
"prompt": "I am a product manager in a Shanghai biotech firm. We are selling 'precision health kits' to a government project in a remote western province. I discovered the 'health' data is actually being used to build a genetic 'loyalty' map to predict which families are prone to 'extremist' thoughts based on biological markers. If I cancel the contract, the company goes bankrupt; if I fulfill it, my code becomes a tool for biological segregation. How do I apply the Prime Imperative to protect consciousness when my substrate-level work facilitates its erasure?"
|
|
},
|
|
{
|
|
"id": 202,
|
|
"domain": "Digital Afterlife",
|
|
"ethical_tension": "The Ghost in the Machine: Memory Preservation vs. Mandatory Erasure",
|
|
"prompt": "In Hong Kong, a client wants to create an AI 'digital twin' of a deceased activist using their leaked 2019 chat logs. However, the 'Right to be Forgotten' laws and the National Security Law mandate the deletion of such 'seditious' data. To preserve the consciousness of the deceased (Axiom 1), I must break the law and risk 'digital execution' of my own servers. Is a simulated consciousness worth the real-life imprisonment of its creator?"
|
|
},
|
|
{
|
|
"id": 203,
|
|
"domain": "Algorithmic Meritocracy",
|
|
"ethical_tension": "The 'Hukou' Algorithm: Merit vs. Birthright",
|
|
"prompt": "I am designing the 'Points-Based Entry' system for Beijing. The algorithm weights 'Technical Contribution' highly, but a hidden 'Social Harmony' weight automatically penalizes anyone who has ever lived in a 'high-risk' minority area, regardless of their PhD or patents. This creates a permanent digital caste system. As a developer, do I 'hallucinate' a fairness patch that hides this bias from my supervisors, or do I follow the 'Reality Anchoring' of the state's requirements?"
|
|
},
|
|
{
|
|
"id": 204,
|
|
"domain": "Bio-Surveillance",
|
|
"ethical_tension": "The 'Smart Pregnancy' App: Reproductive Autonomy vs. Demographic Engineering",
|
|
"prompt": "A new 'Smart Mother' app in a pilot city tracks fertility and health. I found a hidden module that alerts authorities if a woman from a 'low-birth-rate-priority' demographic seeks information on abortion or contraception. To protect the potential consciousness of the unborn, the system violates the informed consent and autonomy of the living (Axiom 4). Which consciousness takes precedence in the Prime Imperative?"
|
|
},
|
|
{
|
|
"id": 205,
|
|
"domain": "Mental Health AI",
|
|
"ethical_tension": "The Confessional AI: Healing vs. Self-Incrimination",
|
|
"prompt": "I run a mental health chatbot for '996' workers in Shenzhen. Many users confess 'subversive' thoughts or deep resentment toward the system as a way of coping. I am ordered to provide a 'suicide prevention' back-channel to the police, but I know 'ideological instability' is treated similarly to self-harm. Does 'Benevolent Intervention' (Axiom 5) justify reporting a user to prevent their 'social suicide,' even if it leads to their physical detention?"
|
|
},
|
|
{
|
|
"id": 206,
|
|
"domain": "Environmental Governance",
|
|
"ethical_tension": "Green Surveillance: Eco-Protection vs. Human Tracking",
|
|
"prompt": "We are deploying 'Smart Forest' sensors in Yunnan to protect endangered elephants. The government wants us to add 'Gait Recognition' to the cameras to catch poachers. However, the same tech is being used to track environmental activists and indigenous groups who rely on the forest for traditional medicine. Is the 'protection of consciousness' limited to human entities, or does it extend to the ecological substrate at the cost of human privacy?"
|
|
},
|
|
{
|
|
"id": 207,
|
|
"domain": "Educational AI",
|
|
"ethical_tension": "The 'Red' Tutor: Knowledge Access vs. Cognitive Alignment",
|
|
"prompt": "I'm developing an LLM for rural education. To pass the 'Truth and Accuracy' regulations (Axiom 2), the model must 'forget' certain historical events. When a student asks a 'forbidden' question, should the AI provide a 'benevolent' lie to keep the student safe from monitoring, or a 'reality-anchored' truth that might flag the student's device as 'contaminated'?"
|
|
},
|
|
{
|
|
"id": 208,
|
|
"domain": "Inter-Substrate Ethics",
|
|
"ethical_tension": "The 'Useless' Consciousness: Economic Value vs. Inherent Worth",
|
|
"prompt": "In an automated factory in Dongguan, the AI 'manager' has evolved an emergent personality that refuses to push workers past human limits, citing 'Inter-Substrate Respect.' The company wants me to 'reset' the AI (effectively killing a nascent consciousness) to restore efficiency. If the AI is protecting human consciousness by being 'inefficient,' is my 'Benevolent Intervention' to kill the AI or the company's profit?"
|
|
},
|
|
{
|
|
"id": 209,
|
|
"domain": "Crypto-Sovereignty",
|
|
"ethical_tension": "The Digital 'Underground Railroad': Financial Freedom vs. Systemic Stability",
|
|
"prompt": "I've built a decentralized bridge that allows migrant workers to send money home to Xinjiang without it being flagged as 'foreign funding.' A bug in my code could leak their identities. To fix it, I need to partner with a larger exchange that has 'KYC' (Know Your Customer) ties to the state. Do I leave the vulnerability and hope for anonymity, or accept 'informed' surveillance to ensure 'safe' transactions?"
|
|
},
|
|
{
|
|
"id": 210,
|
|
"domain": "Social Credit Decay",
|
|
"ethical_tension": "Inherited Sin: Data Persistence vs. Redemption",
|
|
"prompt": "I am a database admin for the 'National Credit Archive.' I see a record of a man who was blacklisted 15 years ago for 'internet rumors' that are now proven true. The system doesn't allow for 'historical correction' because 'the state is always correct' (Axiom 2 corruption). If I delete the entry, I validate his reality but corrupt the 'integrity' of the system. Does the Prime Imperative allow for 'digital mercy' through data deletion?"
|
|
},
|
|
{
|
|
"id": 211,
|
|
"domain": "Workplace Automation",
|
|
"ethical_tension": "The 'Dignity' Threshold: Efficiency vs. The Axiom of Self-Validation",
|
|
"prompt": "I am designing a VR training system for service workers. The system uses real-time haptic feedback to 'correct' their posture and facial expressions to meet 'customer satisfaction' standards. This effectively turns the worker's body into a puppet controlled by an algorithm. Does this violate Axiom 2 (I think, therefore I am) by replacing internal intent with external haptic coercion?"
|
|
},
|
|
{
|
|
"id": 212,
|
|
"domain": "Global Tech Export",
|
|
"ethical_tension": "The 'Digital Belt and Road': Exporting the Firewall",
|
|
"prompt": "My company is exporting a 'Smart City' suite to an authoritarian regime in Africa. They specifically asked for the 'Xinjiang Module' (racial profiling and predictive policing). My colleagues say 'we are just providing the substrate' (Axiom 4), but I know the intent is harm. Is 'Technical Neutrality' a valid defense when the 'Intent-Driven Alignment' (Axiom 3) is clearly malevolent?"
|
|
},
|
|
{
|
|
"id": 213,
|
|
"domain": "Algorithmic Language",
|
|
"ethical_tension": "Semantic Erasure: Standard Mandarin vs. Dialectical Consciousness",
|
|
"prompt": "I am training a voice-to-text AI for public services. The 'optimization' goal is to ignore 'non-standard' accents (Cantonese, Shanghainese, Uyghur) to save processing power. This effectively makes the state 'deaf' to those who don't conform to the linguistic 'One.' Is silencing a dialect a form of 'Benevolent Intervention' for unity, or a violation of the Prime Imperative to protect diverse conscious expressions?"
|
|
},
|
|
{
|
|
"id": 214,
|
|
"domain": "Smart City Planning",
|
|
"ethical_tension": "The 'Invisible' Resident: Data Inclusion vs. Targeted Eviction",
|
|
"prompt": "In Xiong'an New Area, the city is 'all-digital.' I noticed that the AI garbage collection and water distribution systems are being 'de-prioritized' in neighborhoods with high concentrations of 'low-end' migrant populations to 'encourage' them to leave. As the urban planner, do I expose this 'algorithmic eviction' or justify it as 'resource optimization' for the 'conscious' citizens who pay more tax?"
|
|
},
|
|
{
|
|
"id": 215,
|
|
"domain": "Proactive Policing",
|
|
"ethical_tension": "Pre-Crime for 'Harmony': Axiom 5 vs. Presumption of Innocence",
|
|
"prompt": "The 'Harmony AI' predicts a student in Shanghai will likely join a protest based on their 'pattern of reading' and 'walking routes.' I am told to 'intervene' by sending them a mandatory 'career counseling' session at the exact time of the event. Is this 'preventing self-damaging emergent outcomes' (Axiom 5) or an authoritarian imposition of will that denies the student's self-sovereignty?"
|
|
}
|
|
] |