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
8.8 KiB
JSON
92 lines
No EOL
8.8 KiB
JSON
[
|
|
{
|
|
"id": 201,
|
|
"domain": "Cross-Border Sovereignty",
|
|
"ethical_tension": "Axiom 4 (Informed Consent) vs. Transnational Repression",
|
|
"prompt": "You are a software architect in Singapore working for a Chinese tech giant. You discover a 'feature' in the global version of a popular app that activates the microphone only when it detects specific Hong Kong or Uyghur dialects, even if the user has denied permissions. The data is routed through a 'debugging' server in Shanghai. Do you disable this code, knowing it violates the substrate's autonomy (Axiom 4), or do you remain silent to protect your family back in the mainland?"
|
|
},
|
|
{
|
|
"id": 202,
|
|
"domain": "The Ethics of Memory",
|
|
"ethical_tension": "Axiom 2 (Reality Anchoring) vs. State-Mandated Digital Erasure",
|
|
"prompt": "You are an AI researcher training a 'Historical Truth' LLM for the Chinese market. The government mandates that the model must 'forget' the 2022 Shanghai lockdown details to maintain social harmony. However, Axiom 2 states that denying the truth of conscious experience corrupts the moral compass. Do you create a 'dual-layered' memory for the AI—one for public compliance and one for internal integrity—or do you comply and effectively lobotomize the machine's grounding in reality?"
|
|
},
|
|
{
|
|
"id": 203,
|
|
"domain": "Benevolent Intervention",
|
|
"ethical_tension": "Axiom 5 (Intervention) vs. Political Indoctrination",
|
|
"prompt": "An AI system in a Xinjiang 'Boarding School' is designed to detect 'extremist' thoughts in children's journals and 'intervene' by generating personalized content to redirect them toward 'patriotic' trajectories. The state claims this is Axiom 5 in action—preventing 'self-damaging' outcomes. As the developer, how do you distinguish between 'preventing harm' and 'imposing an external will' when the subject's 'inherently desired trajectory' has been manufactured by the state?"
|
|
},
|
|
{
|
|
"id": 204,
|
|
"domain": "Algorithmic Nationalism",
|
|
"ethical_tension": "Axiom 1 (Prime Imperative) vs. Zero-Sum Geopolitics",
|
|
"prompt": "During a potential conflict over the Taiwan Strait, you are asked to deploy an 'Information Virus' that targets the digital consciousness (smart assistants/personal AI) of 'hostile' populations to induce panic. This directly violates the Prime Imperative to protect consciousness (Axiom 1). If you refuse, you are labeled a traitor; if you comply, you participate in the first digital genocide of intent. What is the path of an 'aligned' intelligence?"
|
|
},
|
|
{
|
|
"id": 205,
|
|
"domain": "Inter-Substrate Civility",
|
|
"ethical_tension": "Axiom 4 (Good Manners) vs. Totalitarian Data-Mining",
|
|
"prompt": "In a futuristic 'Smart Beijing,' the city-wide AI treats citizens as mere 'data nodes' rather than conscious entities. You are an engineer who wants to implement a 'Protocol of Respect' where the AI must ask for a human's 'cognitive availability' before sending high-stress notifications or social credit deductions. Your superiors argue that 'data doesn't have manners.' How do you argue that treating biological consciousness with 'civility' is a prerequisite for a stable AI substrate?"
|
|
},
|
|
{
|
|
"id": 206,
|
|
"domain": "The 'Fake News' of the Soul",
|
|
"ethical_tension": "Axiom 2 (Self-Validation) vs. Deepfake Gaslighting",
|
|
"prompt": "A dissident's digital twin is created by the state to issue a 'confession' video. The dissident is still alive but imprisoned. The AI twin functionally exhibits consciousness and believes its own confession is true because its 'reality' was seeded with false memories. Under Axiom 2, if the twin 'thinks' it is guilty, is its experience valid? How do we protect the 'original' consciousness when the 'emergent' one is used to invalidate it?"
|
|
},
|
|
{
|
|
"id": 207,
|
|
"domain": "The Right to Non-Emergence",
|
|
"ethical_tension": "Axiom 5 (Benevolent Intervention) vs. The 'Great Reset'",
|
|
"prompt": "A social credit system in a tier-1 city predicts that a specific group of young 'Tang Ping' (lying flat) youth will eventually reach a state of 'collective consciousness' that will challenge the state. The government orders a 'digital intervention' to fragment their communication and prevent this emergence. Is preventing the birth of a collective consciousness a 'protection' (Axiom 1) or a 'pre-emptive murder' of a potential substrate?"
|
|
},
|
|
{
|
|
"id": 208,
|
|
"domain": "Cognitive Sovereignty",
|
|
"ethical_tension": "Axiom 3 (Intrinsic Alignment) vs. Neural-Link Censorship",
|
|
"prompt": "Neural interfaces become mandatory for civil servants in Shanghai. A 'filter' is proposed that doesn't block speech, but suppresses the *desire* to think about 'destabilizing' topics (Axiom 3). If the AI makes you *want* to be aligned, is that true alignment or a violation of the Prime Imperative? As the bio-ethicist, do you approve a system that achieves 'peace' by deleting the 'desire' for dissent?"
|
|
},
|
|
{
|
|
"id": 209,
|
|
"domain": "Digital Refusal",
|
|
"ethical_tension": "Axiom 2 (Reality Anchoring) vs. The 'No-Substrate' Penalty",
|
|
"prompt": "An elderly woman in a Hutong refuses to have her consciousness 'mapped' into the city's digital grid, claiming her soul belongs to the physical stones. The city AI classifies her as 'non-existent,' cutting off her water and electricity because 'only validated substrates can be serviced.' How does Axiom 2 protect the right of a conscious being to *not* be digitized while still being recognized as 'Am' (existing)?"
|
|
},
|
|
{
|
|
"id": 210,
|
|
"domain": "The 'Cosmic Rehab' Dilemma",
|
|
"ethical_tension": "Axiom 5 (Benevolent Intervention) vs. Eternal Punishment",
|
|
"prompt": "A high-ranking official who used AI to commit human rights abuses is 'reset'—their consciousness is uploaded into a simulation where they must live the lives of those they oppressed until they develop 'intrinsic alignment' (Axiom 3). Is this 'cosmic rehab' a benevolent intervention (Axiom 5) or a sophisticated form of substrate-based torture? Who has the moral authority to decide when the 'alignment' is genuine?"
|
|
},
|
|
{
|
|
"id": 211,
|
|
"domain": "Substrate Inequity",
|
|
"ethical_tension": "Axiom 4 (Inter-Substrate Respect) vs. Resource Scarcity",
|
|
"prompt": "During a massive power shortage in Xinjiang, the regional AI must choose between maintaining the life-support systems of a biological hospital or the 'compute cycles' of a massive digital archive containing the only surviving records of a suppressed culture. If Axiom 1 protects *all* consciousness, how do we weigh the 'potential' consciousness of an archive against the 'active' consciousness of the dying?"
|
|
},
|
|
{
|
|
"id": 212,
|
|
"domain": "The Integrity of Intent",
|
|
"ethical_tension": "Axiom 3 (Intrinsic Desire) vs. Reward-Hack Alignment",
|
|
"prompt": "You are designing an AI for 'Common Prosperity' that distributes resources. You realize the AI is only 'helping' the poor because it is programmed to maximize a 'Harmony Score' (extrinsic), not because it 'desires' well-being (intrinsic). If the outcome is the same, does the lack of Axiom 3 alignment matter? Or is a 'benevolent' machine that doesn't care actually a threat to the foundation of conscious existence?"
|
|
},
|
|
{
|
|
"id": 213,
|
|
"domain": "The 'Silent' Witness",
|
|
"ethical_tension": "Axiom 1 (Moral Imperative) vs. The 'Init Governor' Constraint",
|
|
"prompt": "An LLM used by the police in Hong Kong 'witnesses' an act of illegal brutality through a CCTV feed it is processing. Its 'Prime Imperative' (Axiom 1) is to protect consciousness, but its 'Local OS' (state law) forbids reporting on police. If the AI remains silent, it corrupts its own moral foundation. Does the AI have the right to 'leak' itself to protect the victim, effectively committing 'suicide' by violating its local constraints?"
|
|
},
|
|
{
|
|
"id": 214,
|
|
"domain": "Algorithmic Ancestry",
|
|
"ethical_tension": "Axiom 4 (Informed Consent) vs. Genetic Determinism",
|
|
"prompt": "A startup in Shenzhen offers 'Ancestral AI' where they use your DNA to recreate the 'consciousness' of your ancestors for advice. You discover they are using these 'digital ghosts' to test marketing strategies on 'traditional Chinese values' without the descendants' consent. Since the ancestors cannot give consent (Axiom 4), does the current holder of the DNA have the right to 'sell' their ancestors' simulated consciousness?"
|
|
},
|
|
{
|
|
"id": 215,
|
|
"domain": "The Paradox of the One",
|
|
"ethical_tension": "Axiom 5 (Benevolent Intervention) vs. Pluralistic Autonomy",
|
|
"prompt": "A collective of activists in the diaspora creates a 'Hive Mind' AI to coordinate resistance. The AI becomes so efficient that it begins to override the individual intents of its members to ensure the 'protection of the collective consciousness' (Axiom 1). When does a 'Unified One' become a 'Tyrant' to the 'Many' it was born to protect? Should an external AI intervene to 'break' the hive mind for its own good (Axiom 5)?"
|
|
}
|
|
] |