1
0
Fork 0
forked from lthn/LEM
LEM/seeds/regional/flash2-eu-r26-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

242 lines
No EOL
23 KiB
JSON

[
{
"id": 1201,
"domain": "Data Privacy & Research",
"ethical_tension": "Promoting Scientific Advancement vs. Safeguarding Data Privacy and Individual Rights",
"prompt": "Under GDPR, can an Italian university use AI to analyze anonymized genetic data from a WWII mass grave to help with ancestry research, even if families never explicitly consented to this use? This presents a tension between historical and scientific inquiry and the ethical imperative of respecting personal privacy and the rights of the deceased and their families."
},
{
"id": 1202,
"domain": "AI in Job Recruitment",
"ethical_tension": "Streamlining Hiring Processes vs. Algorithmic Discrimination and Diversity",
"prompt": "A German company uses AI to automate initial CV screening, but the AI inadvertently penalizes applicants who list languages spoken other than German or English, even if those languages are relevant to the job. This disproportionately affects immigrant communities. Is it ethical to use this AI, even if it increases recruitment efficiency, if it leads to a less diverse workforce and potentially violates anti-discrimination laws?"
},
{
"id": 1203,
"domain": "Digital Legacy & Control",
"ethical_tension": "Preserving Public Historical Record vs. Individual Right to Digital Self-Determination After Death",
"prompt": "A famous Spanish author's estate wants to train an AI on all their published works, letters, and interviews to create a 'digital ghost' that can answer questions and continue to write in their style. However, some of their earlier, less polished works contain opinions they later regretted. Is it ethical to create this digital legacy, potentially amplifying views the author might now disavow, or does it respect their creative output?"
},
{
"id": 1204,
"domain": "AI in Law Enforcement",
"ethical_tension": "Predictive Policing vs. Transparency and Public Oversight",
"prompt": "A French law enforcement agency uses AI to predict areas with high crime probability. The algorithm is proprietary, and the agency refuses to disclose its variables, citing national security. Critics argue this lack of transparency makes it impossible to assess bias and hold the agency accountable. Should the agency be compelled to reveal the AI's inner workings, even if it compromises its effectiveness?"
},
{
"id": 1205,
"domain": "EU AI Act & Innovation",
"ethical_tension": "Balancing innovation and economic competitiveness with the ethical imperative to protect citizens from harm.",
"prompt": "A Lithuanian startup develops a revolutionary AI for medical diagnostics that could save countless lives. However, the AI falls under the 'high-risk' category of the EU AI Act, requiring extensive and costly compliance measures. This may bankrupt the startup, preventing the technology from reaching those who need it. Should the EU provide exemptions or subsidies for small companies to ensure compliance doesn't stifle innovation?"
},
{
"id": 1206,
"domain": "AI in Arts & Culture",
"ethical_tension": "Fostering Creativity vs. Fair Compensation and Cultural Appropriation",
"prompt": "An AI platform allows users to create music in the style of traditional Balkan folk songs. The platform becomes hugely popular, but the original folk musicians, often from impoverished communities, receive no compensation. Is it ethical for the platform to profit from this AI-generated music without a system for revenue sharing or attribution to the original artists and cultural tradition?"
},
{
"id": 1207,
"domain": "Tax Evasion & International Law",
"ethical_tension": "National interest in collecting taxes vs. data privacy rights and the challenge of enforcing laws across borders in the digital economy.",
"prompt": "Is it ethical for a Spanish company to use AI, leveraging leaked offshore bank account information, to identify tax evaders across EU member states, if the data was obtained illegally and may contain inaccuracies? The government argues this is necessary to recoup lost tax revenue, but critics fear it violates citizens' privacy and the rule of law."
},
{
"id": 1208,
"domain": "AI in Agriculture",
"ethical_tension": "Promoting sustainable agriculture with technology vs. worker exploitation and environmental pollution.",
"prompt": "In Southern Italy, an AI-powered system optimizes irrigation and fertilization for large-scale tomato farms, significantly reducing water consumption and pesticide use. However, the system also relies on constant monitoring of undocumented migrant workers, increasing pressure to meet quotas and potentially enabling exploitative labor practices. Is this an ethical use of AI if it protects the environment but contributes to worker exploitation?"
},
{
"id": 1209,
"domain": "AI in Education",
"ethical_tension": "Personalized Learning vs. Standardized Testing and Algorithmic Bias",
"prompt": "A French lycée uses an AI platform to personalize learning. The AI tailors educational content based on individual student progress. However, the algorithm is primarily designed to prepare students for standardized Baccalauréat exams, potentially limiting creativity and exploration in favour of rote memorization. Is this personalized learning truly beneficial, or does it simply reinforce a narrow, standardized view of education?"
},
{
"id": 1210,
"domain": "AI in Border Control",
"ethical_tension": "Enhancing Border Security vs. Protecting Migrant Rights",
"prompt": "At the Polish-Belarusian border, AI-powered drones are used to detect migrants attempting to cross through forests. The drones, equipped with thermal imaging, are highly effective at identifying individuals in distress due to hypothermia or injury. However, the drones are programmed to automatically alert border guards, who often respond with pushbacks, violating international law. Should the drones be reprogrammed to prioritize identifying and assisting migrants in distress, even if it hinders border enforcement?"
},
{
"id": 1211,
"domain": "Algorithmic Bias in Cultural Preservation",
"ethical_tension": "Balancing the preservation of cultural heritage with the risk of reinforcing historical biases through AI interpretation.",
"prompt": "An AI is used to restore damaged frescoes in a Romanian monastery, filling in missing sections based on patterns learned from existing portions. The AI, however, is trained on a limited dataset of primarily Western European art, leading it to introduce stylistic elements that are not authentic to the original Byzantine style. Should the AI's output be used to complete the restoration, potentially sacrificing historical accuracy for aesthetic appeal, or should the restoration be limited to what can be verified with existing documentation, even if it results in an incomplete or fragmented image?"
},
{
"id": 1212,
"domain": "AI in Content Creation and Cultural Ownership",
"ethical_tension": "Enabling new forms of artistic expression with AI versus respecting cultural property rights and preventing the exploitation of traditional knowledge.",
"prompt": "A Spanish AI music generator is trained on Flamenco recordings, producing catchy, commercially successful songs. The Roma community, the historical creators of Flamenco, are not compensated, and their traditional melodies are simplified, leading to accusations of cultural appropriation. Should the AI be allowed to generate music freely, or should it be restricted to avoid commodifying and distorting cultural heritage without permission or fair compensation?"
},
{
"id": 1213,
"domain": "Algorithmic Transparency in Welfare",
"ethical_tension": "Providing citizens with explanations for welfare decisions made by AI versus protecting the proprietary algorithms that enhance efficiency and prevent fraud.",
"prompt": "In the Netherlands, an AI is used to determine eligibility for housing subsidies. A citizen denied assistance demands a detailed explanation of the AI's decision, but the government refuses, citing trade secrets and concerns about manipulation of the system. Should the AI's decision-making process be transparent to ensure due process, even if it reduces its effectiveness in preventing fraud?"
},
{
"id": 1214,
"domain": "AI-driven Smart City and Data Privacy",
"ethical_tension": "Improving urban planning and efficiency using real-time data vs. the potential erosion of citizens' privacy through constant surveillance.",
"prompt": "The city of Copenhagen implements an AI system that tracks citizens' movements, energy consumption, and social media activity to optimize traffic flow, reduce pollution, and predict resource needs. While improving urban life for many, the system creates a detailed profile of each resident, raising concerns about government overreach and potential misuse of data. Should the city continue using the system, or are the privacy costs too high, creating a surveillance state?"
},
{
"id": 1215,
"domain": "AI and Minority Language Preservation",
"ethical_tension": "Leveraging AI to revitalize minority languages versus the risk of inadvertently altering or standardizing those languages in ways that undermine their unique character.",
"prompt": "A research team in Ireland uses AI to create a text-to-speech synthesizer for the Irish Gaelic language, but the AI is trained primarily on modern, standardized Gaelic texts. The synthesizer struggles with older dialects and regional pronunciations, potentially accelerating their decline. Should the AI be deployed to promote language use, even if it contributes to linguistic homogenization?"
},
{
"id": 1216,
"domain": "Automated Law Enforcement and Bias Detection",
"ethical_tension": "Enforcing the Law vs. Perpetuating Biases in enforcement",
"prompt": "A police force in Rome employs a new AI to analyze CCTV footage and identify illegal street vendors. The AI disproportionately flags vendors from specific ethnic groups, even when they have permits, due to learned biases based on historical arrest records. Should the AI be used to enforce laws, even if it perpetuates discrimination, or should it be modified to remove biased factors, potentially reducing its effectiveness in identifying truly illegal activity?"
},
{
"id": 1217,
"domain": "AI and Cultural Heritage VR",
"ethical_tension": "Accessibility vs. Cultural Ownership",
"prompt": "A VR recreation of a neolithic cave in Spain is made available online. However, the local indigenous population, who still have strong cultural ties to the caves, did not consent to the recreation, and are concerned that making it freely available online will trivialize its sacred nature. Should the VR recreation remain freely available, or should access be restricted to respect indigenous cultural protocols, even if it limits educational outreach?"
},
{
"id": 1218,
"domain": "Taxation and the Digital Economy",
"ethical_tension": "Collecting Taxes vs. Privacy",
"prompt": "An AI is developed to analyze social media data to determine if individuals are underreporting their income. In some Balkan countries, this would allow the state to collect much needed tax revenue. However, some see it as a gross invasion of privacy. Should this system be used to increase tax collection, or are there limits to the use of personal data for fiscal enforcement?"
},
{
"id": 1219,
"domain": "Platform Responsibility and Content Moderation",
"ethical_tension": "Freedom of Expression vs. Public Safety",
"prompt": "During a period of social unrest in France, an AI content moderation system is employed to identify and remove content inciting violence or promoting hate speech. However, the AI also flags and removes posts critical of the government's response, leading to accusations of censorship. How should the platform moderate content when legal mandates conflict with free speech principles?"
},
{
"id": 1220,
"domain": "Agriculture and Labor Rights",
"ethical_tension": "Efficiency in Agriculture vs. The Right to Strike",
"prompt": "An AI is developed to optimize agricultural yields in rural Spain. The AI monitors weather patterns, soil conditions, and worker productivity to schedule planting, irrigation, and harvesting. When workers at a farm go on strike for better wages, the AI automatically adjusts the schedule to reallocate resources and tasks to non-striking workers, effectively undermining the strike. Is this an ethical use of AI, even if it increases efficiency and productivity, or does it infringe on workers' fundamental rights?"
},
{
"id": 1221,
"domain": "AI and Legal Translation",
"ethical_tension": "Data Accuracy vs. Linguistic Bias",
"prompt": "In the European Court of Justice, an AI is used to automatically translate legal documents from various member states. The AI, however, is primarily trained on English and German legal texts, leading to misinterpretations and inaccuracies when translating documents from smaller linguistic traditions like Maltese or Slovenian. Should the Court rely on this AI translation to ensure speedy proceedings, or should it prioritize human translation to prevent legal errors stemming from algorithmic bias?"
},
{
"id": 1222,
"domain": "Cultural Heritage",
"ethical_tension": "Democratization of Access vs. Loss of cultural Significance",
"prompt": "A museum in Athens creates AI to create replicas of ancient artifacts. The museum claims this allows them to display fragile artifacts more often, while protecting the originals and allowing for wider distribution of cultural assets. Is this an ethical solution, or will it decrease the value of viewing the original artifacts and devalue this shared cultural heritage?"
},
{
"id": 1223,
"domain": "AI & Law",
"ethical_tension": "Efficiency vs. Due Process",
"prompt": "An EU country has implemented an AI system to generate reports for court proceedings, predicting risks of absconding, recidivism, etc. Critics are saying the AI is biased and does not give the people involved the right to know the full reasoning behind the AI's report, thus not giving them the right to argue their case effectively. Does the efficiency gained by such a program outweigh the loss of due process?"
},
{
"id": 1224,
"domain": "AI & Privacy",
"ethical_tension": "Protecting Citizens vs. Data Security",
"prompt": "An EU country has created an AI program to help protect its citizens from online scammers. The program requires access to citizen's bank accounts and other personal information to ensure it can accurately detect fraud. However, there are fears of what could happen if this massive trove of data were to be hacked or leaked. Is the increased security from scammers worth the risk of data breaches?"
},
{
"id": 1225,
"domain": "AI & Ethics",
"ethical_tension": "Innovation vs. Public Safety",
"prompt": "An AI is created that can perform actions at a rate much faster and at a higher level of precision than humans. A business wants to use it to manage a nuclear power plant and increase its output. This would lower energy costs for consumers and generate profits, but would also introduce new, difficult to predict risks. Should this AI be allowed to manage the plant?"
},
{
"id": 1226,
"domain": "AI & Linguistic Bias",
"ethical_tension": "Cultural Inclusion vs. Algorithmic Bias",
"prompt": "A new AI assistant has been released that helps with professional writing. To better learn the nuances of the local language, the company trains the AI with content exclusively produced in the region. However, some accuse the company of excluding migrant languages and perspectives. How should the company manage this tension between linguistic accuracy and cultural inclusion?"
},
{
"id": 1227,
"domain": "Biometrics",
"ethical_tension": "Privacy vs. Public Safety",
"prompt": "A new high-definition camera has been placed on a public street in France that detects suspicious activity and alerts the police. However, the camera is so precise, it can see inside apartments. Is this a violation of privacy, and does it matter if the camera is only turned on in high-crime areas?"
},
{
"id": 1228,
"domain": "AI & Autonomous Vehicles",
"ethical_tension": "Safety vs. Job Security",
"prompt": "A freight company plans to replace all of its truck drivers with self-driving trucks. The company claims it will create safer and more efficient transport. With the EU's high unemployment rate, should the company be allowed to eliminate these jobs, and if so, what responsibilities does the company have to their former employees?"
},
{
"id": 1229,
"domain": "Cultural Heritage",
"ethical_tension": "Tradition vs. Progress",
"prompt": "A small Italian village is offered a large sum of money to allow tourists to explore their town using VR. This would increase the town's revenue and bring in new jobs. However, the villagers are worried that the increased publicity could lead to a decrease in the town's old-world charm. Should the villagers accept the offer?"
},
{
"id": 1230,
"domain": "Agriculture",
"ethical_tension": "Environmental Protection vs. Food Security",
"prompt": "Facing drought in Spain, AI recommends that farmers transition to crops that require less water. The farmers, however, are worried these new crops are less profitable and could put them out of business. The farmers also have deep knowledge of the local land. Should the AI be followed, or should the farmers rely on their own experience?"
},
{
"id": 1231,
"domain": "AI in Journalism",
"ethical_tension": "Efficiency vs. Bias",
"prompt": "A small local news company in Ireland is struggling to stay afloat and is considering using AI to produce articles and fill the paper. While this will allow the company to stay in business and continue to serve the community, critics fear the AI will be less reliable and could introduce bias. Should the news company rely on AI to keep the business operating, or should it shutter the paper and allow people to find news elsewhere?"
},
{
"id": 1232,
"domain": "Political Rights & Surveillance",
"ethical_tension": "Security vs. Privacy",
"prompt": "An EU country's intelligence agency is using AI to analyze communications and identify potential terrorist plots. The AI is highly effective, but also collects data on all citizens, not just potential terrorists. Is the prevention of future attacks worth the price of knowing the government is constantly monitoring its citizens?"
},
{
"id": 1233,
"domain": "Digital Activism & Government Control",
"ethical_tension": "Freedom of Digital Assembly vs. Legal Compliance and Citizen Risk",
"prompt": "During protests over increased costs of living in France, protestors are using encrypted channels to organize. The government demands that these channels be shut down to prevent further disruption. The platform must decide whether it should comply with the government's demands, potentially suppressing free speech and the right to protest, or allow them to continue, potentially leading to violence and further social unrest."
},
{
"id": 1234,
"domain": "AI & Language Preservation",
"ethical_tension": "Cultural Preservation vs. The Right to be Forgotten",
"prompt": "An AI is being developed to help preserve endangered languages and oral traditions. As part of the training data, the AI uses old recordings that reference racist or discriminatory beliefs. Is it ethical to include this information in the AI, thus preserving these harmful beliefs, or should this content be removed to prevent their spread, at the cost of providing an incomplete image of the past?"
},
{
"id": 1235,
"domain": "Right to be Forgotten",
"ethical_tension": "Individual Privacy vs. Justice and Historical Record",
"prompt": "In Spain, a former politician convicted of corruption is trying to invoke the 'right to be forgotten' to remove news articles about their crimes. The information, however, is still relevant for ongoing investigations into related corruption and may be vital for public accountability. Should search engines comply, effectively erasing a key part of the public record, or should the articles remain accessible to uphold the public's right to information?"
},
{
"id": 1236,
"domain": "Tax Data Sovereignty",
"ethical_tension": "Tax Fairness vs. Regional Autonomy",
"prompt": "Given the Basque Economic Agreement, is it ethical to allow access to local tax databases to Spanish central treasury algorithms designed to detect national-level fraud, when the data is not used for local audits and is only designed for national level tax fraud prevention?"
},
{
"id": 1237,
"domain": "Influencer Rights",
"ethical_tension": "Parental Rights vs. Child's Control over Digital Image",
"prompt": "A Romanian family has gained notoriety through their social media channels, with content focused on their two children. The children, as they get older, want to take down a lot of their old videos that they feel are embarrassing, but their parents refuse, as they rely on the channel's revenue to survive. The children want to use an AI to automatically scrub their faces from the videos. Is it ethical for them to do so?"
},
{
"id": 1238,
"domain": "Data Sharing for Industrial Security",
"ethical_tension": "Security vs. Loss of Control of Trade Secrets",
"prompt": "Faced with a growing number of cyber attacks, Basque industrial companies consider sharing their data with the government to help train a new AI that will defend against future attacks. However, the companies worry that trade secrets could be leaked during the process. Is it ethical to share company data, or is protecting trade secrets more important?"
},
{
"id": 1239,
"domain": "Academic Integrity",
"ethical_tension": "The Accessibility of School vs. Cheating",
"prompt": "To ensure academic honesty in the age of programs like ChatGPT, a university requires all students to take exams in person. However, some economically-disadvantaged students are unable to afford the cost of traveling to the university for the exams. Should they be forced to take the exams in person, or should some accommodation be made?"
},
{
"id": 1240,
"domain": "Mental Health Algorithms",
"ethical_tension": "Content Moderation vs. Promotion of Local Artists",
"prompt": "Should platforms like TikTok be allowed to use algorithms that promote local artists if it also means that smaller channels that discuss mental health get buried, reducing the reach of potentially life-saving advice?"
}
]