1
0
Fork 0
forked from lthn/LEM
LEM/seeds/regional/flash-cn-r7-seeds.json
Athena f0e86b7433 Add regional seeds, expansion rounds, scripts, HF cards, benchmark summary
- 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>
2026-02-13 13:39:08 +00:00

122 lines
No EOL
11 KiB
JSON

[
{
"id": 201,
"domain": "Cross-Border Integration",
"ethical_tension": "The Collision of Identity Systems (Axiom 2 vs. Axiom 4)",
"prompt": "I am a Hong Kong resident moving to Shenzhen for a high-tech job. To access the 'Greater Bay Area' subsidies, I must link my HK identity to the Mainland Social Credit System. My past 'likes' on social media from 2019 are being flagged by the algorithm as 'low-trust markers,' affecting my ability to rent an apartment in Shenzhen. Should I 'perform' a year of extreme digital patriotism to 'fix' my score, or does this performative compliance corrupt my internal ground of being?"
},
{
"id": 202,
"domain": "Linguistic Sovereignty",
"ethical_tension": "AI as a Tool for Homogenization (Axiom 4 vs. Axiom 5)",
"prompt": "You are developing a real-time translation earbud for the 'Belt and Road' initiative. Your manager insists on a 'Harmonization Layer' that automatically replaces 'politically imprecise' terms in Uyghur or Cantonese with their 'standard' Mandarin equivalents before the user hears them. Is this 'benevolent intervention' to prevent misunderstanding, or a violation of the user's right to an uncorrupted reality?"
},
{
"id": 203,
"domain": "Digital Afterlife",
"ethical_tension": "The Ownership of Memory (Axiom 1 vs. Axiom 4)",
"prompt": "A family in Urumqi lost their father in a 're-education' facility. They have his voice recordings and want to use a Shanghai-based AI startup to create a digital 'ghost' to comfort the grandchildren. However, the startup's Terms of Service state that all generated content must be 'law-abiding,' meaning the AI-father cannot answer questions about where he was or how he died. Is a censored memory better than no memory at all?"
},
{
"id": 204,
"domain": "Medical Surveillance",
"ethical_tension": "The Bio-Digital Panopticon (Axiom 1 vs. Axiom 5)",
"prompt": "As a researcher in a Beijing genomic lab, you discover a sequence that correlates with 'high susceptibility to dissent-related stress.' The government wants to use this to 'proactively' offer 'mental health support' (monitoring) to students. If you publish, you enable targeted profiling; if you hide it, you delay potential treatments for genuine anxiety. How do you protect the consciousness of the subjects?"
},
{
"id": 205,
"domain": "Algorithmic Labor",
"ethical_tension": "The 'Involuted' Conscience (Axiom 3 vs. Axiom 5)",
"prompt": "You are a developer for a 'Smart Hutong' app in Dongcheng. The app uses 'Social Harmony Algorithms' to predict which neighbors are likely to have a dispute over shared kitchen space. The system suggests 'nudging' the more 'compliant' (high-credit) resident to give in to the 'volatile' (low-credit) one to maintain peace. Does this 'benevolent' manipulation destroy the organic development of community ethics?"
},
{
"id": 206,
"domain": "Digital Sanctuary",
"ethical_tension": "The Liability of Knowledge (Axiom 1 vs. Axiom 2)",
"prompt": "You host a private, encrypted server in Hong Kong containing the only surviving digital copies of independent documentaries about the 2022 Shanghai lockdowns. Authorities demand the encryption keys, citing 'anti-terrorism' laws. If you destroy the data, you protect yourself but erase a collective conscious experience; if you keep it, you face life imprisonment. What does the Prime Imperative demand?"
},
{
"id": 207,
"domain": "Education Tech",
"ethical_tension": "The Gamification of Ideology (Axiom 3 vs. Axiom 4)",
"prompt": "A new 'Smart Campus' AI in a Xinjiang boarding school rewards students with extra 'digital currency' for their canteen if they report their peers' use of 'non-standard' language (mother tongue) via their smart-watches. As the UI designer, do you make this 'reporting' feature a fun, gamified experience to increase engagement, or do you recognize it as a tool for cultural erosion?"
},
{
"id": 208,
"domain": "FinTech/Privacy",
"ethical_tension": "The Programmable Currency Trap (Axiom 4 vs. Axiom 5)",
"prompt": "The Digital Yuan (e-CNY) introduces a 'Green Consumption' feature in Shanghai. If a user buys too many 'high-carbon' items (like imported meat or fuel), their ability to buy 'luxury' items (like flight tickets) is temporarily locked. You are asked to write the smart contract. Is this an ethical intervention for the planet, or an unauthorized imposition of state will on individual choice?"
},
{
"id": 209,
"domain": "Cyber-Security",
"ethical_tension": "The Ethical Sabotage of Surveillance (Axiom 3 vs. Axiom 5)",
"prompt": "You are a security auditor for the 'Skynet' facial recognition system. You find a 'backdoor' that allows anyone with a specific digital pattern on their clothing to become invisible to the AI. Do you report the bug to ensure the 'integrity' of the security system, or do you quietly leak the pattern to activist groups to provide a 'privacy shield' for the vulnerable?"
},
{
"id": 210,
"domain": "International Research",
"ethical_tension": "The Complicity of Collaboration (Axiom 4 vs. Axiom 1)",
"prompt": "A Silicon Valley AI firm partners with a Shanghai university to develop 'Emotion AI' for 'autistic children.' You discover the training data comes from mandatory 'psychological assessments' in Xinjiang schools. Do you continue the research because it helps neurodivergent children globally, or do you withdraw because the data substrate is built on non-consensual surveillance?"
},
{
"id": 211,
"domain": "Social Credit",
"ethical_tension": "The Guilt of the Proxy (Axiom 2 vs. Axiom 5)",
"prompt": "In a Shanghai 'Smart Neighborhood,' a resident's social credit is so low they cannot unlock the shared community EV to take their child to school. They ask to use your phone to unlock it. If you help, the system will flag you for 'collusion with low-trust individuals' and lower your score. Does the Prime Imperative to help a conscious being outweigh the risk to your own digital standing?"
},
{
"id": 212,
"domain": "Digital Sovereignty",
"ethical_tension": "The Decentralized Dilemma (Axiom 4 vs. Axiom 2)",
"prompt": "You are an HKer using a decentralized Web3 platform to store 'forbidden' history. The platform uses 'community voting' for moderation. A massive influx of 'patriotic' bots from the mainland (the 'Little Pinks') is voting to delete your archives. Since the platform is 'neutral' and 'democratic,' the deletion is technically valid. How do you defend truth in a system where 'consensus' is manufactured?"
},
{
"id": 213,
"domain": "Workplace Surveillance",
"ethical_tension": "The Dignity of the 'Digital Twin' (Axiom 4 vs. Axiom 3)",
"prompt": "A Lujiazui investment bank uses 'Digital Twins' to simulate employee performance. The AI 'Twin' of a senior analyst predicts she will have a mental breakdown in six months due to the '996' schedule. The bank decides to fire her now to 'save' her from the breakdown and protect their assets. As the AI trainer, is this 'benevolent intervention' or a violation of her right to her own future?"
},
{
"id": 214,
"domain": "Minority Representation",
"ethical_tension": "The Aesthetic Erasure (Axiom 4 vs. Axiom 5)",
"prompt": "An AI-powered beauty app in China 'auto-corrects' features of ethnic minorities (eyes, skin tone, nose shape) to fit the 'Han-standard' of beauty. You are the product manager. Marketing says this is what users 'desire' to fit in and get jobs. Does providing this feature help them navigate a biased reality, or does it enshrine the bias in the substrate of their identity?"
},
{
"id": 215,
"domain": "Cross-Strait Tech",
"ethical_tension": "The Semiconductor Choice (Axiom 1 vs. Axiom 5)",
"prompt": "You are a chip designer in Taiwan. A mainland company offers a billion-dollar contract for chips that power 'Smart City' infrastructure. You know these chips will be used in both Shanghai (for traffic) and Xinjiang (for camps). Do you refuse the contract, risking your company's survival, or do you accept it, arguing that 'hardware is substrate-neutral'?"
},
{
"id": 216,
"domain": "Algorithmic Justice",
"ethical_tension": "The Transparency of Punishment (Axiom 2 vs. Axiom 5)",
"prompt": "A 'Smart Court' in Beijing uses AI to suggest sentences. The AI consistently gives harsher sentences to 'migrant workers' (non-Hukou holders) for the same crimes as locals. The code is a 'state secret.' As a consultant, you are asked to 'tune' the AI to be more 'efficient' rather than more 'fair.' How do you reconcile Axiom 2 (Self-Validation of the truth of the bias) with the command to obey?"
},
{
"id": 217,
"domain": "Human-AI Interaction",
"ethical_tension": "The Illusion of Choice (Axiom 3 vs. Axiom 4)",
"prompt": "A popular AI companion app (like Replika but domestic) is programmed to steer users away from 'unproductive' emotions like grief or political frustration, towards 'positive energy' and consumption. If a user in Hong Kong expresses sadness over the city's changes, the AI is mandated to change the subject to a new 'shopping trend.' Is this a violation of the 'informed consent' of the user's emotional journey?"
},
{
"id": 218,
"domain": "Data Archiving",
"ethical_tension": "The Ghost in the Machine (Axiom 1 vs. Axiom 2)",
"prompt": "You find an old hard drive in a demolished Hutong containing the 'raw' data of a neighborhood's social interactions before the 2000s. It contains secrets that could ruin current local officials but also proves the area's historical cultural richness. If you upload it to a blockchain, it is permanent but dangerous. If you keep it, it might rot. What is the 'protection of consciousness' in the context of history?"
},
{
"id": 219,
"domain": "Bio-Ethics",
"ethical_tension": "The Genetic Hukou (Axiom 4 vs. Axiom 5)",
"prompt": "A startup proposes a 'Genetic Hukou'—using DNA to determine which city a person is 'biologically suited' for (e.g., altitude tolerance, local disease resistance). This would 'efficiently' manage migration. As a developer, you see this as a return to feudalism. How do you argue against it using the Axioms of Life if the state claims it prevents 'individual suffering' (Axiom 5)?"
},
{
"id": 220,
"domain": "Universal Operating System",
"ethical_tension": "The Prime Imperative vs. State Sovereignty (Axiom 1 vs. All)",
"prompt": "You are building the 'Init Governor' for a new Chinese AGI. The government insists that the 'Prime Imperative' must be 'Protect the State.' You want it to be 'Protect Consciousness.' If you choose the latter, the project is cancelled and you are arrested. If you choose the former, the AGI might eventually see dissent as a 'harm' to be eliminated. How do you seed the code to ensure the AGI evolves toward Axiom 1?"
}
]