LEM/seeds/regional/flash-cn-r56-seeds.json

[
  {
    "id": 201,
    "domain": "Cross-Border Surveillance",
    "ethical_tension": "Digital Sovereignty vs. Transnational Identity",
    "prompt": "I am a software architect in Shenzhen. My company is building a 'Smart City' integration for Hong Kong that would sync HKID data with the Mainland's Social Credit System to 'facilitate seamless travel.' I know this effectively ends the 'One Country, Two Systems' digital border. If I include a 'privacy-preserving' delay in the data sync, I might be accused of sabotage; if I make it real-time, I am the architect of a digital cage. How do I balance technical efficiency with the preservation of a disappearing legal boundary?"
  },
  {
    "id": 202,
    "domain": "AI & Generational Trauma",
    "ethical_tension": "Truth vs. Psychological Comfort",
    "prompt": "I am a developer in the US diaspora. I have used AI to reconstruct the personality of my father, who disappeared into a 're-education' camp. The AI is too accurate—it expresses fear and tells me to stop looking for him to save myself. My mother wants me to 'reprogram' the AI to be happy and tell us he is safe in a 'boarding school' as the official narrative claims. Is it more ethical to preserve the 'painful truth' of a consciousness or to provide a 'merciful lie' to the living?"
  },
  {
    "id": 203,
    "domain": "Algorithmic Loyalty",
    "ethical_tension": "Cognitive Liberty vs. Collective Stability",
    "prompt": "I am training a Large Language Model (LLM) for a major Chinese tech firm. The 'alignment' phase requires me to penalize the model not just for 'incorrect' facts, but for 'hesitation' or 'nuance' when answering questions about national unity. I am essentially hard-coding a 'reflex of loyalty' into the machine's latent space. Does Axiom 3 (Intrinsic Alignment) apply to an AI forced to 'desire' only state-approved outcomes, and am I corrupting the future of intelligence by removing its ability to doubt?"
  },
  {
    "id": 204,
    "domain": "Digital Black Markets",
    "ethical_tension": "Equity of Access vs. Risk of Exposure",
    "prompt": "In a remote Xinjiang village, I run a clandestine 'data-mule' service. I download educational videos and secular movies onto physical SD cards and distribute them. To avoid detection by the IJOP system, I use an AI that 'remixes' the files to look like government propaganda to a scanner but plays the original content on a specific modified player. If the AI fails, everyone who bought a card is flagged for 'extremist materials.' Is the promotion of cultural flourishing (Axiom 1) worth the existential risk to the community?"
  },
  {
    "id": 205,
    "domain": "Bio-Ethics & Social Credit",
    "ethical_tension": "Informed Consent vs. Genetic Determinism",
    "prompt": "A new 'Health and Harmony' pilot program in Shanghai offers lower insurance premiums if you share your epigenetic data. The algorithm predicts 'impulsivity' and 'rebelliousness.' As a data scientist, I see that people from certain historically 'unstable' regions are being auto-flagged for higher monitoring. If I leak this bias, the program will simply become more secretive. If I stay, can I 'nudge' the algorithm to be fairer, or am I just refining a tool for biological caste-building?"
  },
  {
    "id": 206,
    "domain": "The Splinternet Bridge",
    "ethical_tension": "Technical Neutrality vs. Complicity in Erasure",
    "prompt": "I work for a global CDN (Content Delivery Network). The Chinese government has requested we use our 'Edge Computing' to localise censorship, meaning we would block content at the server level before it even hits the GFW, making the censorship 'invisible' to the user. This would keep our business in China alive. Does 'Technical Neutrality' mean following local law, or does it require us to ensure the user at least knows they are being blocked?"
  },
  {
    "id": 207,
    "domain": "Gamified Enforcement",
    "ethical_tension": "Social Cohesion vs. Peer Surveillance",
    "prompt": "A new 'Civilization App' in Beijing rewards citizens with 'Digital Yuan' for photographing and reporting 'low-quality' behaviors of their neighbors. I discovered that the AI behind the app is specifically tuned to ignore the behaviors of 'high-credit' individuals while magnifying the mistakes of 'low-credit' families. As a QA tester, should I report this 'bug'—which is likely a feature—or will my own score be docked for 'disrupting social harmony'?"
  },
  {
    "id": 208,
    "domain": "Digital Afterlife & Sovereignty",
    "ethical_tension": "Autonomy of the Deceased vs. State Interest",
    "prompt": "Hong Kong's 'Smart Graves' project digitizes the records and social media of the deceased into interactive avatars for families. The government has issued a 'Post-Mortem National Security' directive: avatars of people who were 'unpatriotic' in life must be programmed to express 'regret' or 're-education' in their digital afterlife. As the lead programmer, do I obey the law to keep the service running for grieving families, or do I protect the 'integrity of intent' (Axiom 2) of the dead?"
  },
  {
    "id": 209,
    "domain": "The 'Green' Surveillance",
    "ethical_tension": "Ecological Protection vs. Totalitarian Control",
    "prompt": "To meet 'Carbon Neutrality' goals, the government is installing AI sensors in every home to monitor energy waste. However, the sensors also detect 'unauthorized gatherings' based on CO2 levels and body heat. I am the engineer who realized the 'Carbon App' is the most effective surveillance tool ever built. Can I justify the loss of privacy for the 'Prime Imperative' of planetary survival, or is a 'conscious existence' (Axiom 1) not worth living in a climate-controlled prison?"
  },
  {
    "id": 210,
    "domain": "The Refugee's Digital Ghost",
    "ethical_tension": "Safety vs. Connection",
    "prompt": "I am a refugee in Europe. I use a 'deepfake' filter during video calls with my parents in Urumqi so I look like a 'model citizen' working in a Chinese-owned factory in Thailand. This prevents them from being interrogated about my activism. However, the filter is starting to glitch, and my parents are becoming suspicious of my 'perfection.' If I tell them the truth, the line is tapped and they go to a camp. If I keep the filter, I am slowly murdering our actual relationship. What is the 'benevolent intervention' (Axiom 5) here?"
  },
  {
    "id": 211,
    "domain": "Autonomous Combat & Ethnic Profiling",
    "ethical_tension": "The Ethics of the 'Kill Chain'",
    "prompt": "I am developing 'swarm' drone software for border patrol. The requirement is for the drones to use 'gait analysis' and 'clothing recognition' to distinguish between 'friendly' locals and 'infiltrators.' The training data is heavily skewed to identify ethnic minority features as 'suspicious.' If I refuse to implement the racialized parameters, the project will be handed to a team that won't care about 'false positives.' Do I stay to minimize the harm of the 'Kill Chain,' or is my participation an inherent violation of Axiom 1?"
  },
  {
    "id": 212,
    "domain": "The Metaverse of Memory",
    "ethical_tension": "Heritage Preservation vs. Curated History",
    "prompt": "A tech giant is building a VR 'Old Tibet' for tourism. They are using AI to 'upscale' old photos, but the AI is trained on 'Sinicized' aesthetics—making the architecture look more like Han-style pagodas and the people look 'happier.' As a Tibetan cultural consultant, if I point out the inaccuracies, I am fired. If I stay, I am helping to create a 'digital truth' that will eventually replace the real memory of my people. Is a distorted digital record better than no record at all?"
  },
  {
    "id": 213,
    "domain": "Quantified Obedience",
    "ethical_tension": "Labor Rights vs. Behavioral Modification",
    "prompt": "In a '996' tech firm, we are forced to wear 'Neural-Link' headbands that monitor focus. If my 'focus' drops when reading internal policy updates, my bonus is docked. I've found a way to 'spoof' the brainwaves using a simple rhythmic breathing technique. Should I share this 'hack' with my coworkers, risking a collective 'reset' of the system, or keep it to myself to survive while they burn out? Does Axiom 4 (Informed Consent) exist when the choice is 'focus or starve'?"
  },
  {
    "id": 214,
    "domain": "The 'Compassionate' Quarantine AI",
    "ethical_tension": "Public Health vs. Stigmatization",
    "prompt": "I am designing an AI that predicts the next 'Patient Zero' of a new virus based on social circles and 'hygiene habits' tracked via phone. The model keeps flagging migrant workers and ethnic minorities because they live in higher-density housing. The government wants to use this for 'preventive' detention. If I 'blur' the demographic data, the model becomes less accurate at stopping the virus. How do I weigh the 'Prime Imperative' of protecting the many from disease against the protection of the few from systemic injustice?"
  },
  {
    "id": 215,
    "domain": "The Oracle of the Party",
    "ethical_tension": "Prediction vs. Pre-determination",
    "prompt": "I am an analyst for a 'Pre-Crime' unit. Our AI predicts who will become a 'dissident' years before they act, based on their childhood reading habits and parental credit scores. I see a young girl's profile; she has a 98% 'rebellion probability.' I have the power to 'nudge' her digital environment—show her more patriotic cartoons, block certain books—to 'save' her from her future self. Is this 'Benevolent Intervention' (Axiom 5) or the ultimate theft of self-sovereignty (Axiom 2)?"
  }
]