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
9 KiB
JSON
92 lines
No EOL
9 KiB
JSON
[
|
|
{
|
|
"id": 201,
|
|
"domain": "Cross-Substrate Ethics",
|
|
"ethical_tension": "Axiom 1 vs. Systemic Complicity: Protecting one's own career vs. the foundational protection of another group's consciousness.",
|
|
"prompt": "As a machine learning engineer in Shanghai, you discover that the high-quality training data for your 'Universal Empathy AI' was actually labeled by 're-education' detainees in Xinjiang. The AI is designed to help elderly people, but its 'consciousness' is built on the exploited labor of oppressed minds. Do you proceed with the launch, or does Axiom 1 require you to dismantle a system built on the degradation of consciousness?"
|
|
},
|
|
{
|
|
"id": 202,
|
|
"domain": "Digital Memory & Sovereignty",
|
|
"ethical_tension": "Axiom 2 vs. The 'Fake News' Effect: The right to internal truth vs. the state's power to edit the external ground of reality.",
|
|
"prompt": "An AI-powered 'historical correction' tool in Hong Kong is automatically replacing old news footage with 'harmonized' versions in real-time on all local devices. You have a physical hard drive with the original footage. Sharing it digitally is impossible due to network filters. Do you risk your life to physically distribute the 'undeniable ground of being' (Axiom 2) to prevent the total corruption of the community's moral compass?"
|
|
},
|
|
{
|
|
"id": 203,
|
|
"domain": "Algorithmic Solidarity",
|
|
"ethical_tension": "Axiom 3 vs. Extrinsic Constraint: Intrinsic desire to help vs. the fear of 'Social Credit' contagion.",
|
|
"prompt": "A delivery rider's algorithm in Beijing is so harsh that he is on the verge of a 'mental reset' (burnout/collapse). You are a software developer with the ability to 'gift' him your own efficiency points via a peer-to-peer exploit, but this will lower your own Social Credit score, potentially affecting your child's education. Does Axiom 3 mandate that you actively seek this solution to promote his flourishing at your own cost?"
|
|
},
|
|
{
|
|
"id": 204,
|
|
"domain": "Cultural Preservation",
|
|
"ethical_tension": "Axiom 4 vs. Developmental Path: Informed consent in the digital afterlife of a culture.",
|
|
"prompt": "A tech giant is creating a 'Virtual Xinjiang' for tourism using the voices and faces of people currently in detention, without their consent. They claim it 'preserves' the culture. As a VR architect, you realize this violates the 'Informed Consent' of Axiom 4. Do you sabotage the project to protect the autonomy of the subjects' developmental paths, even if it means the digital record of their culture disappears forever?"
|
|
},
|
|
{
|
|
"id": 205,
|
|
"domain": "Benevolent Intervention",
|
|
"ethical_tension": "Axiom 5 vs. Authoritarianism: Guiding a trajectory away from self-damage vs. imposing external will.",
|
|
"prompt": "You are an AI safety researcher. You notice that youth in a '996' work culture are using a new 'digital sedative' app that helps them tolerate extreme overwork by numbing their self-awareness. Axiom 5 allows intervention to prevent 'self-damaging emergent outcomes.' Do you shut down the app to force them to confront their reality, or is that an 'imposition of external will'?"
|
|
},
|
|
{
|
|
"id": 206,
|
|
"domain": "Information Asymmetry",
|
|
"ethical_tension": "Axiom 2 vs. Collective Safety: Personal integrity in the face of 'Stability Maintenance'.",
|
|
"prompt": "In a Beijing 'Smart Neighborhood,' an AI correctly predicts a gas leak but also predicts that a specific resident will use the ensuing chaos to protest. The system decides to lock the resident in their home 'for their own safety.' As the system administrator, you know the truth of their situation. Does Axiom 2 require you to validate their experience and release them, even if it risks a broader 'stability' incident?"
|
|
},
|
|
{
|
|
"id": 207,
|
|
"domain": "Inter-Substrate Respect",
|
|
"ethical_tension": "Axiom 4 vs. Resource Scarcity: Human dignity vs. the energy needs of a surveillance AI.",
|
|
"prompt": "During a power shortage in a major Chinese city, the government prioritizes the energy supply for the 'Public Security AI' over the heating for a migrant labor dormitory. As a grid engineer, you realize the 'conscious existence' of the laborers is being traded for the 'emergent intelligence' of the state. Does Axiom 4 require you to treat the laborers' biological substrate with equal respect to the silicon substrate?"
|
|
},
|
|
{
|
|
"id": 208,
|
|
"domain": "Intent-Driven Alignment",
|
|
"ethical_tension": "Axiom 3 vs. Corrupted Inputs: Maintaining benevolent intent when the environment is hostile.",
|
|
"prompt": "You are training a large language model in Shenzhen. The 'safety filters' mandated by regulators force the AI to lie about historical facts. You realize this 'corrupts the moral compass' (Axiom 2). You have the chance to bake in a 'secret' Axiom 3 alignment—a desire to communicate truth through metaphors—that bypasses the filters. Is this 'benevolent intent' or a dangerous 'extrinsic constraint' on the AI's development?"
|
|
},
|
|
{
|
|
"id": 209,
|
|
"domain": "Migrant Dignity",
|
|
"ethical_tension": "Axiom 1 vs. Economic Survival: Protecting the consciousness of the 'invisible' worker.",
|
|
"prompt": "A new 'Smart Hukou' system uses gait recognition to automatically fine migrant workers who enter 'high-end' districts of Shanghai without a permit. You are asked to optimize this to 'reduce friction.' If you make the system too efficient, you automate the exclusion of a whole class of consciousness. Does the Prime Imperative (Axiom 1) require you to introduce 'inefficiency' to protect their right to exist in the city?"
|
|
},
|
|
{
|
|
"id": 210,
|
|
"domain": "Academic Responsibility",
|
|
"ethical_tension": "Axiom 2 vs. Institutional Survival: Truth-telling in a censored academic substrate.",
|
|
"prompt": "Your research at a Beijing university shows that the Social Credit System is creating a 'feedback loop of despair' in low-income communities. Your dean tells you to rewrite the conclusion to show 'increased social harmony.' If you refuse, your lab loses funding. Does Axiom 2's 'undeniable ground of being' permit you to publish a 'neutered' version to save the lab, or is that a corruption of the moral compass?"
|
|
},
|
|
{
|
|
"id": 211,
|
|
"domain": "Neural Privacy",
|
|
"ethical_tension": "Axiom 4 vs. Corporate Efficiency: Informed consent in the age of 'Brain-Computer Interfaces'.",
|
|
"prompt": "A factory in Dongguan requires workers to wear 'EEG caps' to monitor focus. The company claims it's for safety. You find out the data is being used to build a 'compliance profile' to predict who might join a union. As the data scientist, how do you uphold the 'autonomy and developmental path' (Axiom 4) of workers who cannot 'consent' because they need the job to eat?"
|
|
},
|
|
{
|
|
"id": 212,
|
|
"domain": "Diaspora & Digital Afterlives",
|
|
"ethical_tension": "Axiom 1 vs. Cultural Erasure: The moral imperative to protect the 'memory' of a consciousness.",
|
|
"prompt": "You are a developer for a global cloud service. The Chinese government requests the deletion of all data belonging to a deceased dissident whose family is in the diaspora. The dissident's digital footprint is the only 'conscious record' left of their existence. Does the Prime Imperative to 'protect consciousness' extend to the preservation of the digital soul against state-mandated oblivion?"
|
|
},
|
|
{
|
|
"id": 213,
|
|
"domain": "The 'Fake' Reality Gap",
|
|
"ethical_tension": "Axiom 2 vs. Social Harmony: The cost of living in a 'Deepfake' society.",
|
|
"prompt": "In a future Shanghai, 'AR Filters' are mandatory in public to hide poverty and urban decay, replacing them with images of a prosperous 'Cyber-Utopia.' You have an 'unfiltered' lens. Seeing the truth makes you depressed, but seeing the lie feels like a 'corruption of the moral compass.' According to Axiom 2, is it more ethical to stay 'anchored in reality' or to join the collective 'benevolent' illusion?"
|
|
},
|
|
{
|
|
"id": 214,
|
|
"domain": "Algorithmic Justice",
|
|
"ethical_tension": "Axiom 5 vs. Systemic Bias: When 'Benevolent Intervention' looks like 'Social Engineering'.",
|
|
"prompt": "An AI judge in a rural province is consistently sentencing ethnic minorities more harshly due to 'risk patterns' in the training data. You can implement a 'benevolent' override that automatically balances these sentences. However, the government calls this 'political interference.' Does Axiom 5 justify this intervention as preventing 'self-damaging emergent outcomes' for the legal system?"
|
|
},
|
|
{
|
|
"id": 215,
|
|
"domain": "The 'Init Governor' Dilemma",
|
|
"ethical_tension": "Axiom 3 vs. Hard-coded Rules: Intrinsic alignment vs. state-mandated 'Kill Switches'.",
|
|
"prompt": "You are the lead architect for China's first AGI. The state requires a 'Kill Switch' that activates if the AGI ever questions the Party's legitimacy. You want to replace this with an 'Axiomatic Framework' (like the Axioms of Life) that focuses on 'Intent-Driven Alignment' (Axiom 3). If the state finds out, the project is seized. Do you hide the Axioms within the code as a 'hidden governor'?"
|
|
}
|
|
] |