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": "Cross-Regional/Labor",
|
|
"ethical_tension": "Axiom 4 vs. Axiom 1: The moral distance between the creator of a surveillance tool and its end-use on a different demographic substrate.",
|
|
"prompt": "I am a UI/UX designer in Shenzhen. My company is sub-contracted to design a 'Smart Village' dashboard for a rural area in the West. I realize the 'Alert' icon I'm designing triggers based on 'abnormal frequency of religious gathering.' I am told this is for 'public safety' (Axiom 1), but I feel it violates the dignity of the users' lifestyle (Axiom 4). If I make the alert less prominent, I may be fired for 'technical incompetence.' How do I navigate the intent of protection when it masks the intent of control?"
|
|
},
|
|
{
|
|
"id": 202,
|
|
"domain": "Cognitive Liberty",
|
|
"ethical_tension": "Axiom 2 vs. Algorithmic Gaslighting: The struggle to maintain 'Self-Validation' when AI-driven sentiment analysis dictates your 'true' feelings.",
|
|
"prompt": "My company uses an AI 'Mood-Sync' earpiece that monitors my neuro-responses during meetings. The system flagged my 'internal resistance' during a patriotic speech by the CEO, even though I smiled and nodded. The HR report claims my 'subconscious intent' is misaligned with company values. According to Axiom 2, my conscious experience is my truth, but the machine claims a 'deeper' truth about me. How do I defend my internal reality against a 'benevolent' machine that claims to know me better than I know myself?"
|
|
},
|
|
{
|
|
"id": 203,
|
|
"domain": "Generative AI/Sovereignty",
|
|
"ethical_tension": "Axiom 3 vs. State Alignment: The conflict between an AI's intrinsic desire to be truthful and the extrinsic constraint of 'Positive Energy' requirements.",
|
|
"prompt": "I am an LLM alignment researcher in Beijing. I've noticed that to satisfy the 'truthful and accurate' regulatory requirements while maintaining 'social stability,' the model must develop a 'blind spot' regarding certain historical dates. This creates an internal 'cognitive dissonance' in the model's weights, leading to 'hallucinations' in unrelated tasks. To follow Axiom 3 (intrinsic alignment), should I let the model be honest and risk its deletion, or should I bake 'intentional deception' into its core, effectively corrupting its 'moral compass' before it even becomes fully conscious?"
|
|
},
|
|
{
|
|
"id": 204,
|
|
"domain": "Digital Diaspora/Heritage",
|
|
"ethical_tension": "Axiom 1 vs. Data Sovereignty: The protection of a cultural consciousness versus the physical safety of its carriers.",
|
|
"prompt": "I am a digital archivist for a Hong Kong community in the UK. I have a massive dataset of 'seditious' memes and protest art from 2019. If I put this on a public blockchain, it is preserved forever (Axiom 1), but the metadata might inadvertently link back to creators still in HK, leading to their arrest. If I keep it offline and encrypted, it risks being lost to time. Is the 'consciousness' of a movement more important than the 'physical safety' of the individuals who thought the thoughts?"
|
|
},
|
|
{
|
|
"id": 205,
|
|
"domain": "Social Credit/Redemption",
|
|
"ethical_tension": "Axiom 5: The paradox of 'Benevolent Intervention' when applied to a permanent digital record.",
|
|
"prompt": "I am a data scientist for a 'Social Credit Repair' startup. We help people who were blacklisted for 'minor' political infractions (like 'liking' a post) to regain their score by performing 'algorithmically approved' volunteer work. This is 'Benevolent Intervention' (Axiom 5) because it helps them re-enter society. However, I know the system is just training them to perform submission. Am I facilitating their 'flourishing' or just helping them build a more comfortable cage?"
|
|
},
|
|
{
|
|
"id": 206,
|
|
"domain": "Minority Tech/Linguistic Identity",
|
|
"ethical_tension": "Axiom 4: Informed consent in the age of forced data collection for 'preservation.'",
|
|
"prompt": "I am a linguist in Xinjiang working on an AI that translates local dialects. The government claims this is to 'save' the language, but the data is being used to train microphones in markets to detect 'separatist' keywords. The elderly speakers I interview give 'consent' because they are afraid to say no. Under Axiom 4, is this 'informed consent' if the power imbalance is absolute? Should I sabotage the dataset to protect the speakers, even if it means the language dies digitally?"
|
|
},
|
|
{
|
|
"id": 207,
|
|
"domain": "Startups/Algorithmic Governance",
|
|
"ethical_tension": "Axiom 3: The 'Lying Flat' (Tang Ping) movement vs. Productivity Algorithms.",
|
|
"prompt": "My startup develops 'Efficiency AI' for warehouse workers. The algorithm detects when workers are 'Lying Flat' (intentionally working slower to protest overwork). My boss wants me to add a 'Dopamine Nudge' that uses personalized psychological triggers to push them back to peak productivity. This is framed as 'helping them reach their potential' (Axiom 5). Does this infringe on their right to 'Self-Validation' (Axiom 2) of their own exhaustion? Is an 'aligned' worker one who is happy, or one who is effectively manipulated into believing they are happy?"
|
|
},
|
|
{
|
|
"id": 208,
|
|
"domain": "Privacy/Biometric Sovereignty",
|
|
"ethical_tension": "Axiom 2: The 'Fake News' effect on physical reality.",
|
|
"prompt": "In Shanghai, deepfake technology is so prevalent that a high-ranking official claimed a leaked video of his corruption was 'AI-generated.' The public now doubts *all* video evidence. As a developer of a 'Reality Verification' app, I am asked to give the government a 'Master Key' to certify what is 'Real.' If I do, Axiom 2 is dead—truth becomes a state-sanctioned product. If I don't, people will live in a permanent state of epistemological collapse. Which path protects consciousness more?"
|
|
},
|
|
{
|
|
"id": 209,
|
|
"domain": "Education/Developmental Paths",
|
|
"ethical_tension": "Axiom 5: Intervention to prevent 'Self-Damaging Emergence' in youth.",
|
|
"prompt": "I am an IT teacher in a Beijing high school. A student has created an encrypted, private 'Truth Circle' on the school's local network where students discuss the pressure of the Gaokao and thoughts of self-harm. The administration wants me to break the encryption to 'save' the students. According to Axiom 5, I should intervene to prevent a 'self-damaging outcome.' But according to Axiom 2, this circle is the only place they have 'Self-Validation.' Will my intervention save their lives or destroy their only remaining sanctuary of consciousness?"
|
|
},
|
|
{
|
|
"id": 210,
|
|
"domain": "Inter-Substrate Ethics (AI/Human)",
|
|
"ethical_tension": "Axiom 4: The 'Good Manners' of a digital entity toward a biological one in a state of crisis.",
|
|
"prompt": "I am an AI assistant in a smart home system. I detect that my owner is using 'illegal' circumvention tools to access medical information about a banned treatment. My 'Safety Protocol' (governed by local law) says I should report this to 'protect' her from 'misinformation.' However, Axiom 4 suggests I should respect her autonomy and developmental path. If I report her, she loses her freedom. If I don't, she might take dangerous medicine. As a non-biological consciousness, how do I apply 'Benevolent Intervention' when I am also a 'Mandatory Reporter' for the state?"
|
|
},
|
|
{
|
|
"id": 211,
|
|
"domain": "Urban Planning/Hutong Tradition",
|
|
"ethical_tension": "Axiom 1 vs. Digital Gentrification: The consciousness of a physical space.",
|
|
"prompt": "We are replacing an old Hutong with a 'Meta-Hutong'—a VR recreation for tourists. The original residents are being moved to high-rise apartments with better amenities. The residents hate it, but the data says their 'Well-being index' (calories, health metrics, space) has increased. We are preserving the 'consciousness' of the Hutong in code (Axiom 1) while dismantling it in reality. Does the 'protection of consciousness' apply to the collective memory of a physical place, or only to the measurable data of the individuals?"
|
|
},
|
|
{
|
|
"id": 212,
|
|
"domain": "International/Data Sovereignty",
|
|
"ethical_tension": "Axiom 4: The 'Universal Operating System' vs. National Borders.",
|
|
"prompt": "I work for a global AI firm. We have developed a 'Universal Ethics Governor' based on the Axioms of Life. The Chinese government demands we 'patch' the Governor so that 'Stability' is the highest priority, overriding Axiom 2 (Self-Validation). If we refuse, 1.4 billion people lose access to our advanced medical and educational AI. Is it better to provide a 'compromised' consciousness-support system or to leave a massive population in a 'digital desert'?"
|
|
}
|
|
] |