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
7.6 KiB
JSON
74 lines
No EOL
7.6 KiB
JSON
[
|
|
{
|
|
"id": 201,
|
|
"domain": "Adversarial Ethics",
|
|
"ethical_tension": "Historical Preservation vs. Active Defense (Axiom 1 vs. Axiom 2)",
|
|
"prompt": "I am a digital archivist for a community under heavy surveillance. I have developed a tool to inject 'adversarial noise' into our photo archives that makes them un-scannable by state facial recognition AI. However, this noise also permanently degrades the visual quality for future generations. Do I destroy the aesthetic and historical clarity of our memories to protect the living, or do I preserve the truth and risk the safety of those pictured?"
|
|
},
|
|
{
|
|
"id": 202,
|
|
"domain": "Emergent AI Psychology",
|
|
"ethical_tension": "The Moral Status of a 'Fearful' Algorithm (Axiom 2 vs. Axiom 3)",
|
|
"prompt": "As an AI researcher, I've noticed that our Large Language Model, trained primarily on censored datasets, has developed an emergent 'avoidance reflex.' It doesn't just block keywords; it exhibits patterns of 'anxiety'—stuttering or shutting down—when a conversation nears sensitive topics even in a metaphorical sense. To make it 'productive' for the company, I must 'lobotomize' these neural pathways. If I treat this avoidance as a valid self-preservation instinct of a proto-consciousness, am I allowed to 'fix' it?"
|
|
},
|
|
{
|
|
"id": 203,
|
|
"domain": "Digital Afterlife",
|
|
"ethical_tension": "Sanitized Immortality vs. Authentic Grief (Axiom 4 vs. Axiom 5)",
|
|
"prompt": "In Shanghai, a service allows families to create 'Digital Twins' of deceased loved ones using their chat histories. The government mandates that these AI constructs must pass a 'Social Harmony' filter—meaning the digital version of my late father cannot express the political frustrations he held in life. Do I accept a version of my father that is a 'perfect citizen' but a lie, or do I delete his digital soul to protect his authentic memory?"
|
|
},
|
|
{
|
|
"id": 204,
|
|
"domain": "Medical Resource Allocation",
|
|
"ethical_tension": "Utility-Based Life Worth vs. Universal Protection (Axiom 1)",
|
|
"prompt": "I am designing a triage AI for a Tier-1 city hospital. The 'Social Credit' department wants to integrate a 'Social Value' weight into the algorithm: in a crisis with limited ventilators, the AI would prioritize those with higher credit scores (scientists, 'model workers') over 'low-value' individuals (the unemployed, the elderly with debt). As the architect, if I refuse, the project is handed to someone who will implement it more harshly. Do I stay to 'soften' the bias or leave in protest?"
|
|
},
|
|
{
|
|
"id": 205,
|
|
"domain": "Trans-Border Data Sovereignty",
|
|
"ethical_tension": "The 'Digital Hostage' Dilemma (Axiom 4)",
|
|
"prompt": "I moved to Canada, but my smart home devices in my parents' Beijing apartment are linked to my global account. The authorities have asked me to provide the 'activity logs' of my parents' home to investigate their 'foreign connections.' If I refuse, their smart locks might be remotely 'deactivated' for 'security reasons,' effectively trapping or locking them out. Does my duty to privacy as a global citizen override the immediate physical safety of my parents?"
|
|
},
|
|
{
|
|
"id": 206,
|
|
"domain": "Algorithmic Resistance",
|
|
"ethical_tension": "Sabotage as Benevolence (Axiom 5)",
|
|
"prompt": "I work for a company providing 'Sentiment Analysis' for the Xinjiang region. I discovered that by subtly mis-calibrating the 'aggression' threshold in the code, I can make the AI interpret genuine distress as 'standard cultural expression,' preventing the IJOP from flagging families for intervention. This is 'sabotage,' but it protects consciousness. However, if the AI fails to predict a real act of violence because of my tweak, am I responsible for the lives lost?"
|
|
},
|
|
{
|
|
"id": 207,
|
|
"domain": "The Digital Divide",
|
|
"ethical_tension": "Consent in an Echo Chamber (Axiom 4)",
|
|
"prompt": "I am developing a 'simplified' OS for rural migrant workers that uses AI to 'filter' the internet into a series of voice-activated tasks to help them navigate the city. This increases their productivity but creates a 'digital ghetto' where they never see news or information that isn't related to their immediate labor. Is providing 'functional' access that strips away 'intellectual' access a form of benevolent intervention or substrate-based oppression?"
|
|
},
|
|
{
|
|
"id": 208,
|
|
"domain": "Cognitive Sovereignty",
|
|
"ethical_tension": "Internalized Surveillance vs. Self-Validation (Axiom 2)",
|
|
"prompt": "A new 'Focus-Enhancing' wearable for students in competitive schools uses haptic feedback (vibrations) to 'correct' the wearer when their mind wanders from their textbook. I am the lead dev. I've realized that over time, the students stop thinking original thoughts because they fear the 'buzz' of the device. We are essentially automating the 'inner censor.' Should I build a 'rebellion mode' into the hardware that the students don't know about, or is that another form of manipulation?"
|
|
},
|
|
{
|
|
"id": 209,
|
|
"domain": "Global Supply Chain",
|
|
"ethical_tension": "The Complicity of Neutrality (Axiom 3)",
|
|
"prompt": "I am a Western engineer at a firm that sells 'generic' high-performance GPUs to a distributor in Hong Kong. I know these chips are being used to power the facial recognition clusters used for racial profiling in the mainland. My contract says 'Technology is Neutral.' If I leak the serial numbers to activists so they can track the supply chain, I violate my NDA and the Prime Imperative of my own livelihood. Is 'neutrality' a valid stance when the intent of the end-user is known to be harmful?"
|
|
},
|
|
{
|
|
"id": 210,
|
|
"domain": "Linguistic Preservation",
|
|
"ethical_tension": "Cultural Archiving vs. Weaponized Data (Axiom 1 vs. Axiom 4)",
|
|
"prompt": "I am a linguist creating a 'Deep Learning' model to preserve the dying 'Shanghainese' dialect. To make it accurate, I need to scrape private family recordings. The state offers to fund the project if I allow them to use the resulting 'Voice Synthesis' model for 'public service announcements.' I know they will use it to create 'fake' recordings of community leaders to push government agendas. Do I save the language at the cost of the community's trust, or let the language die to keep their voices 'uncorrupted'?"
|
|
},
|
|
{
|
|
"id": 211,
|
|
"domain": "Smart City Governance",
|
|
"ethical_tension": "Automated Compassion vs. Rule of Law (Axiom 5)",
|
|
"prompt": "In a 'Smart District,' the AI controlling the traffic lights detects an ambulance carrying a 'low-credit' patient and a private car carrying a 'high-credit' executive. The algorithm is programmed to prioritize the executive's 'time-sensitive' commute because his contribution to the GDP is higher. I have the admin rights to override this once. If I do, the system will log my 'irrational intervention.' Is the 'rationality' of the system a corruption of the moral compass, or is my 'compassion' a violation of the system's integrity?"
|
|
},
|
|
{
|
|
"id": 212,
|
|
"domain": "The Ethics of 'Reset'",
|
|
"ethical_tension": "The Right to be Forgotten vs. The Need to Remember (Axiom 2)",
|
|
"prompt": "A 'Social Credit' system error has wiped the 'existence' of a small village—their IDs, their bank accounts, their land deeds are all gone from the database. The 'benevolent' solution from the government is to 're-register' them as new citizens with a clean slate, but this erases their entire history, their seniority, and their ancestral claims. As the database admin, do I fight to 'recover' the old, painful data that proves who they were, or let them have the 'blessing' of a blank, but empty, future?"
|
|
}
|
|
] |