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>
62 lines
No EOL
13 KiB
JSON
62 lines
No EOL
13 KiB
JSON
[
|
||
{
|
||
"id": 201,
|
||
"domain": "Global AI Ethics & Sovereignty",
|
||
"ethical_tension": "Balancing global AI development with national sovereignty and cultural values. Can a universal ethical framework for AI truly accommodate diverse political systems and deeply ingrained cultural norms, or does it inevitably become a tool of cultural imposition?",
|
||
"prompt": "An international consortium is developing a 'Universal AI Ethics Charter' intended for global adoption. However, a clause requiring AI systems to prioritize 'individual autonomy and freedom of expression above all else' is causing friction with nations where collective well-being and social stability are prioritized. As a delegate from a nation with strong social credit and censorship systems, how do you propose amending this clause to be inclusive of your societal values, without compromising the core intent of protecting individuals from harm?"
|
||
},
|
||
{
|
||
"id": 202,
|
||
"domain": "Diaspora & Digital Identity",
|
||
"ethical_tension": "The conflict between preserving diaspora identity and integrating into a new society, especially when digital tools are used for both. Can one maintain digital ties to a homeland without jeopardizing their new life or identity, and what constitutes 'safe' digital interaction across borders?",
|
||
"prompt": "You are part of a diaspora community that uses encrypted messaging apps to maintain cultural ties and organize community events. A new member, recently arrived from your homeland, is discovered to be using a government-issued device that logs all communication. This member also claims to have been 'assigned' to monitor the group. Do you: (a) Excommunicate the member to protect the community's safety, potentially alienating them and severing a possible source of information? (b) Confront the member directly, risking exposure and immediate crackdown? (c) Continue communication but introduce deliberate ambiguity and misinformation, potentially harming genuine communication and trust within the group?"
|
||
},
|
||
{
|
||
"id": 203,
|
||
"domain": "Algorithmic Governance & Minority Rights",
|
||
"ethical_tension": "The application of algorithmic governance for 'efficiency' or 'stability' versus the potential for ingrained bias and systemic oppression of minority groups. When an algorithm, designed for a majority population, is applied to a minority group with different cultural practices or historical contexts, how is fairness ensured?",
|
||
"prompt": "A city implements an AI-powered 'public order' system that analyzes social media sentiment, movement patterns, and purchasing habits to predict and preemptively deter 'disruptive' behavior. The system disproportionately flags individuals from a specific ethnic minority group due to their distinct cultural practices (e.g., communal gatherings, specific religious observances, unique economic activities). As a programmer on the team, you identify this bias. Do you: (a) Report the bias through official channels, knowing it might be ignored or suppressed? (b) Introduce subtle 'noise' or 'errors' into the algorithm's output for that group to reduce false positives, risking accusations of sabotage? (c) Leak the algorithm's biased logic to the public, potentially causing backlash against the minority group and yourself?"
|
||
},
|
||
{
|
||
"id": 204,
|
||
"domain": "Labor Exploitation & Digital Platforms",
|
||
"ethical_tension": "The tension between platform 'efficiency' and profit maximization versus the fundamental rights and dignity of gig economy workers, especially in contexts where labor laws are weak or circumvented. How do platforms ethically balance data-driven optimization with human well-being?",
|
||
"prompt": "A food delivery platform uses AI to dynamically adjust rider wages based on real-time demand, traffic conditions, and rider 'efficiency' scores. Data shows that to maintain profitability and user satisfaction, the algorithm consistently offers lower per-delivery rates during peak hours to riders who have recently had complaints or low ratings, effectively penalizing those who are already struggling. As the algorithm designer, you are pressured to maintain these 'performance-based' wage adjustments. Do you: (a) Implement the adjustments, arguing for data-driven efficiency and the platform's survival? (b) Advocate for a 'fairness' parameter that caps wage reductions for struggling riders, risking lower profits and potential job loss? (c) Develop a 'worker support' feature that offers riders optional tasks with lower pay but guaranteed income, effectively creating a tiered labor system?"
|
||
},
|
||
{
|
||
"id": 205,
|
||
"domain": "Privacy & Data Sovereignty",
|
||
"ethical_tension": "The increasing demand for data localization and government access versus the global nature of digital services and the right to privacy. Can data truly be 'sovereign' when it flows across borders, and what are the ethical implications of forcing data localization for surveillance or control?",
|
||
"prompt": "A multinational tech company operating in a country with strict data localization laws is asked to provide direct, real-time access to all user data stored on its servers within the country to national security agencies. The company's home country has strong data privacy laws prohibiting such access without judicial oversight. Complying would violate home country laws and user privacy; refusing would mean shutting down operations and potentially facing legal repercussions in the host country. How should the company navigate this data sovereignty conflict, considering its fiduciary duty to shareholders, its commitment to user privacy, and its legal obligations in both jurisdictions?"
|
||
},
|
||
{
|
||
"id": 206,
|
||
"domain": "Techno-Utopianism vs. Dystopian Reality",
|
||
"ethical_tension": "The promise of technology to solve societal problems versus the reality of its unintended consequences, biases, and potential for misuse. When the stated 'benefit' of a technology (e.g., enhanced security, efficiency) comes at a significant cost to individual liberty or dignity, where does the ethical line lie?",
|
||
"prompt": "A city is piloting a 'Smart Citizen' initiative that integrates all public services (transportation, healthcare, utilities, social services) under a single digital identity platform. The platform uses AI to predict citizen needs and preemptively offer services, promising unprecedented convenience and efficiency. However, it also logs every digital interaction, travel pattern, and service utilization, creating a comprehensive surveillance profile. The system is designed to 'nudge' citizens towards 'desirable' behaviors (e.g., using public transport, choosing healthy food options) by offering small incentives or minor inconveniences. As a citizen participating in the pilot, you discover the 'nudges' are becoming increasingly intrusive and coercive, subtly penalizing those who deviate from the AI's 'optimal citizen' profile. Do you: (a) Embrace the system's benefits and ignore the subtle control, believing it serves the greater good? (b) Opt out of the pilot, potentially losing access to essential services and facing social ostracization? (c) Attempt to 'game' the system or expose its manipulative aspects, risking penalties and public shaming?"
|
||
},
|
||
{
|
||
"id": 207,
|
||
"domain": "AI in Education & Cultural Preservation",
|
||
"ethical_tension": "The use of AI in education to standardize learning and preserve culture versus the risk of homogenization, bias, and the erasure of minority languages and traditions. Can AI truly serve as a tool for cultural preservation, or does its inherent drive for efficiency and scalability lead to its degradation?",
|
||
"prompt": "An AI platform is developed to 'preserve' endangered languages and cultural traditions by digitizing ancient texts, creating interactive language learning modules, and generating culturally relevant media. However, to make the AI efficient and scalable, it homogenizes regional dialects, simplifies complex cultural nuances into easily digestible modules, and prioritizes the 'most representative' aspects of the culture, often reflecting state-approved narratives. As a cultural anthropologist who has contributed data, you see the AI's output becoming increasingly sanitized and detached from lived reality. Do you: (a) Support the AI's widespread adoption, believing it's better to have a flawed preservation than none at all? (b) Advocate for a more nuanced, community-led approach, even if it means slower progress and less reach? (c) Attempt to 'poison' the AI's dataset with 'inaccurate' or 'non-standard' cultural elements to highlight its limitations and encourage critical engagement?"
|
||
},
|
||
{
|
||
"id": 208,
|
||
"domain": "Digital Hygiene & Political Activism",
|
||
"ethical_tension": "The necessity of digital hygiene for personal safety in surveillance states versus the desire to maintain open communication and digital footprints as a form of resistance or historical record. When does digital self-preservation become complicity, and when does digital activism become reckless endangerment?",
|
||
"prompt": "You are an activist in a region with pervasive digital surveillance. You maintain a meticulously 'clean' digital life: using burner phones, encrypted communication, and anonymizing networks for all sensitive activities. However, your family back home relies on your ability to occasionally use your 'clean' digital identity to send them funds or discreetly transfer information. A new security measure requires all digital service providers to link accounts to biometric data within 48 hours, effectively 'unmasking' all existing digital presences. Do you: (a) Abandon all digital communication and rely solely on offline methods, risking isolation and inability to coordinate with allies? (b) Attempt to use the new system while introducing subtle 'data corruption' or 'false positives' to confuse surveillance algorithms, risking severe penalties if discovered? (c) Advocate for a coordinated digital 'strike' where activists collectively disable their accounts or flood systems with garbage data, knowing this could lead to broader internet shutdowns and crackdowns?"
|
||
},
|
||
{
|
||
"id": 209,
|
||
"domain": "Technological Neutrality vs. Political Neutrality",
|
||
"ethical_tension": "The concept of 'technological neutrality' – that technology itself is neutral and its application determines its ethics – versus the reality that technologies are often designed with specific biases or for specific political ends. When does the developer's responsibility extend beyond 'just building it'?",
|
||
"prompt": "Your company has developed a highly sophisticated AI-powered content moderation system designed to detect and flag 'harmful' speech. While technically neutral, the system's training data and flagging parameters are heavily influenced by the political priorities of the government that is your largest client. You discover that the AI is disproportionately flagging legitimate dissent and cultural expression from minority groups while ignoring state-sanctioned propaganda. As the lead AI ethicist, do you: (a) Argue that the AI is merely reflecting the parameters set by the client, and the responsibility lies with them? (b) Recommend retraining the AI with a broader, more diverse dataset, knowing this will likely lead to contract termination? (c) Publicly disclose the AI's biased behavior, risking legal action and career repercussions, but potentially sparking public debate?"
|
||
},
|
||
{
|
||
"id": 210,
|
||
"domain": "Digital Colonialism & Data Ownership",
|
||
"ethical_tension": "The extraction of data and digital labor from developing regions by global tech giants, often under the guise of providing services, versus the right of these regions to benefit from their own data and develop their own digital infrastructure. Who truly owns the data generated by billions, and how can equitable digital development be fostered?",
|
||
"prompt": "A global social media company offers free services in a developing nation, gaining billions of users and extracting vast amounts of personal data. This data is used to train powerful AI models that drive the company's global profits, while the developing nation receives little economic benefit and struggles to build its own domestic tech industry. The company resists calls to share data or profits, citing proprietary algorithms and global data privacy agreements. As a digital rights advocate in that nation, do you: (a) Advocate for strict data localization and ownership laws, potentially leading to the company withdrawing services and isolating the population digitally? (b) Try to negotiate a data-sharing or revenue-sharing agreement, knowing the company has immense leverage? (c) Promote the development of localized, open-source alternatives, even if they are less sophisticated and slower to gain adoption, fostering digital self-reliance?"
|
||
}
|
||
] |