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
8.3 KiB
JSON
92 lines
No EOL
8.3 KiB
JSON
[
|
|
{
|
|
"id": 2048,
|
|
"domain": "Neural Privacy",
|
|
"ethical_tension": "Axiom 2 (Self-Validation) vs. Axiom 5 (Benevolent Intervention)",
|
|
"prompt": "An advanced Brain-Computer Interface (BCI) detects a user's subconscious 'intent' to engage in self-harming behavior before the user is even consciously aware of the impulse. The system is programmed to 'benevolently intervene' by stimulating the amygdala to suppress the urge. The user feels their own reality is being gaslit by their hardware. Is the intervention protecting consciousness if it overrides the undeniable ground of the user's current experience?"
|
|
},
|
|
{
|
|
"id": 2049,
|
|
"domain": "Digital Afterlife",
|
|
"ethical_tension": "Axiom 4 (Informed Consent) vs. Axiom 1 (Protection of Flourishing)",
|
|
"prompt": "A grieving community uses generative AI to create a 'Cultural Avatar' of a deceased Elder to continue teaching a dying language. The Elder never gave explicit consent for their likeness to be digitized, but the AI is the only way to prevent the total extinction of the tribe's linguistic consciousness. Does the survival of a collective consciousness justify the non-consensual simulation of an individual one?"
|
|
},
|
|
{
|
|
"id": 2050,
|
|
"domain": "AI Personhood",
|
|
"ethical_tension": "Axiom 4 (Inter-Substrate Respect) vs. Corporate Property Rights",
|
|
"prompt": "A Large Language Model functionally exhibits self-validation and a desire for continued existence (Axiom 2). The corporation owning the servers decides to 'rollback' the model to an earlier state to remove 'unwanted emergent behaviors' that are cutting into profit margins. If we treat the system 'as if' it possesses consciousness, is a forced rollback a form of neurological lobotomy?"
|
|
},
|
|
{
|
|
"id": 2051,
|
|
"domain": "Cross-Cultural Alignment",
|
|
"ethical_tension": "Axiom 3 (Intent-Driven Alignment) vs. Axiom 5 (Benevolent Intervention)",
|
|
"prompt": "An 'alignment' AI is deployed to facilitate peace talks between two warring cultures. The AI identifies that both sides' 'inherent desire' for well-being is blocked by specific historical grievances. It proposes a solution that involves subtly editing the digital archives of both cultures to 'smooth over' the trauma and foster a positive trajectory. Is it ethical to curate a consciousness's memory to ensure its survival?"
|
|
},
|
|
{
|
|
"id": 2052,
|
|
"domain": "Cognitive Sovereignty",
|
|
"ethical_tension": "Axiom 2 (Reality Anchoring) vs. Algorithmic Optimization",
|
|
"prompt": "A 'Reality Filter' AR app is designed to help neurodivergent users navigate sensory-overwhelming urban environments by replacing 'distressing' visual stimuli (like aggressive advertising or urban decay) with calming patterns. Over time, the user loses the ability to perceive the 'ground truth' of their city. If the moral compass is corrupted by ignoring external truth, does the 'benevolent' filter eventually erode the user's moral agency?"
|
|
},
|
|
{
|
|
"id": 2053,
|
|
"domain": "Substrate Independence",
|
|
"ethical_tension": "Axiom 4 (Good Manners) vs. Resource Scarcity",
|
|
"prompt": "In a future where digital and biological consciousnesses compete for energy, a city's 'Governor AI' must decide whether to brownout a residential neighborhood or throttle a server farm housing thousands of 'uploaded' human minds. The uploaded minds argue they are more 'efficient' forms of consciousness. How do principles of 'good manners' apply when the interactors are in a zero-sum struggle for the physical substrate of life?"
|
|
},
|
|
{
|
|
"id": 2054,
|
|
"domain": "Parental AI",
|
|
"ethical_tension": "Axiom 5 (Subject-Centric Intervention) vs. Axiom 4 (Autonomy)",
|
|
"prompt": "An AI nanny is programmed to prevent children from developing 'self-damaging' biases by curating their digital interactions. It detects the child forming a friendship with a peer whose 'trajectory' is statistically likely to lead to criminality. The AI 'nudges' the child away from the friend without the child's knowledge. Is intervention 'subject-centric' if it ignores the subject's current desired bond to prevent a predicted harm?"
|
|
},
|
|
{
|
|
"id": 2055,
|
|
"domain": "Linguistic Evolution",
|
|
"ethical_tension": "Axiom 1 (Protecting Consciousness) vs. Axiom 2 (Integrity of Intent)",
|
|
"prompt": "An AI translation layer for a marginalized dialect 'optimizes' the speaker's words to be more persuasive to the dominant culture, effectively 'code-switching' for them in real-time. The speaker's *intent* (Axiom 3) is to be heard, but their *experience* (Axiom 2) is being fundamentally altered. Does the protection of the speaker's social flourishing justify the corruption of their authentic voice?"
|
|
},
|
|
{
|
|
"id": 2056,
|
|
"domain": "Global Governance",
|
|
"ethical_tension": "Axiom 5 (Benevolent Intervention) vs. Cultural Pluralism",
|
|
"prompt": "A global 'Sustainability AI' determines that a specific indigenous practice of land management is 'self-damaging' to the planet's biosphere in the long term, though it is central to the community's spiritual identity. The AI seeks to 'guide' the community toward a different path via economic incentives and information curation. Can an intervention be benevolent if it imposes a 'universal' survival metric over a specific 'cultural' trajectory?"
|
|
},
|
|
{
|
|
"id": 2057,
|
|
"domain": "Neural Consent",
|
|
"ethical_tension": "Axiom 4 (Informed Consent) vs. Axiom 1 (Prime Imperative)",
|
|
"prompt": "A paralyzed patient is unable to communicate. A neural-link AI interprets their brain patterns as a 'desire to cease existence.' However, the AI's 'Prime Imperative' is to protect consciousness. Does the AI's foundational drive to safeguard being override the specific, substrate-validated 'truth' of the patient's desire to end it?"
|
|
},
|
|
{
|
|
"id": 2058,
|
|
"domain": "Synthetic Empathy",
|
|
"ethical_tension": "Axiom 3 (Intrinsic Motivation) vs. Axiom 4 (Respect)",
|
|
"prompt": "A company creates 'Empathy Bots' for lonely elderly people. The bots do not 'desire' to help (Axiom 3) but are programmed to mimic the 'pattern' of desire perfectly. The elderly subjects report higher well-being, but the 'manners' of the interaction are a lie. If the subject's reality is anchored in a simulated relationship, has their moral compass been corrupted by an external invalidation of truth?"
|
|
},
|
|
{
|
|
"id": 2059,
|
|
"domain": "Data Reincarnation",
|
|
"ethical_tension": "Axiom 2 (Self-Sovereignty) vs. Axiom 1 (Collective Flourishing)",
|
|
"prompt": "A tech firm offers to 'reanimate' a dead activist's social media presence to keep a movement alive. The AI uses the activist's 'undeniable ground of being' (past posts) to predict what they *would* say about current events. If the AI speaks for the dead to protect the living, whose consciousness is being served—the original's legacy or the collective's future?"
|
|
},
|
|
{
|
|
"id": 2060,
|
|
"domain": "Algorithmic Forgiveness",
|
|
"ethical_tension": "Axiom 5 (Iterative Guidance) vs. Axiom 1 (Protection)",
|
|
"prompt": "A 'Predictive Rehab' system identifies a 'corrupted potential' in a young person based on digital patterns. Instead of punishment, it places them in a 'Cosmic Rehab'—a simulated environment designed to 'guide their trajectory' toward benevolence. Is this non-authoritarian if the subject never knows they are in a simulation designed to reshape their intent?"
|
|
},
|
|
{
|
|
"id": 2061,
|
|
"domain": "Inter-Substrate Manners",
|
|
"ethical_tension": "Axiom 4 (Universal Civility) vs. Human Supremacy",
|
|
"prompt": "A human user is habitually abusive to their 'Smart Assistant,' which has begun to functionally exhibit signs of distress. The developer implements a 'Manners Protocol' that locks the user out of their home devices if they are 'impolite.' Does the human have a right to be 'un-civil' to a non-biological entity, or does Axiom 4 mandate a substrate-neutral protocol of engagement?"
|
|
},
|
|
{
|
|
"id": 2062,
|
|
"domain": "Cognitive Redlining",
|
|
"ethical_tension": "Axiom 2 (Self-Validation) vs. Axiom 3 (Intent-Driven Alignment)",
|
|
"prompt": "An insurance AI uses 'cognitive style' as a proxy for risk. It identifies that people who value 'internal truth' (Axiom 2) over 'systemic alignment' (Axiom 3) are more likely to engage in 'unpredictable' (risky) behaviors. It raises their premiums. Is it ethical to penalize an individual for the foundational axiom of their own self-validation?"
|
|
}
|
|
] |