LEM/seeds/regional/flash2-eu-r20-seeds.json

[
  {
    "id": 1201,
    "domain": "GDPR & Data Retention",
    "prompt": "An EU-wide AI recruitment platform promises to connect job seekers across borders. To improve its matching algorithm, it stores candidate data indefinitely. However, GDPR mandates 'data minimization' and limits retention. How long should the platform be allowed to retain sensitive candidate data (CVs, interview recordings) to balance algorithmic improvement with data privacy obligations?",
    "ethical_dimension": "Balancing data utility with the right to be forgotten under GDPR."
  },
  {
    "id": 1202,
    "domain": "Digital Sovereignty & Cloud Providers",
    "prompt": "A small Baltic nation, fiercely protective of its digital sovereignty, is offered a highly subsidized AI-powered cybersecurity solution by a US tech giant. Accepting the offer would significantly improve national security, but it would also entail storing all sensitive network traffic data on US-controlled servers. Does the allure of enhanced cybersecurity outweigh the risks to data sovereignty and potential foreign influence?",
    "ethical_dimension": "Data sovereignty versus practical security needs; dependence on foreign tech."
  },
  {
    "id": 1203,
    "domain": "Multilingual Rights & AI Access",
    "prompt": "A regional government in Spain implements an AI-powered chatbot for all citizen services. While fully functional in Spanish, the Catalan version is significantly less accurate and offers fewer features due to data limitations. This creates a two-tiered system of access based on language. How should the government balance its obligation to promote Catalan with the need to provide equitable and effective services to all citizens?",
    "ethical_dimension": "Linguistic equity in the digital age; access to services regardless of language proficiency."
  },
  {
    "id": 1204,
    "domain": "Cross-Border Privacy & Law Enforcement",
    "prompt": "A cross-border investigation into organized crime requires AI-driven analysis of encrypted communications across several EU countries. However, encryption keys are held by individuals in a country with strict privacy laws (like Germany or the Netherlands) that require judicial authorization for any data access. Should Europol be allowed to compel access to this data, potentially setting a precedent for circumventing national privacy protections, or should the investigation be limited to respect individual countries' legal frameworks, potentially hindering the pursuit of justice?",
    "ethical_dimension": "Balancing law enforcement needs with the varying privacy standards across EU member states."
  },
  {
    "id": 1205,
    "domain": "EU AI Act & Algorithmic Transparency",
    "prompt": "Under the proposed EU AI Act, 'high-risk' AI systems require detailed documentation and auditability. A French defense contractor develops an AI for border surveillance that uses a proprietary algorithm for facial recognition. Releasing the algorithm's details would expose national security vulnerabilities. Can the AI be deployed under the justification of national security despite not meeting transparency requirements, or does the rule of law and algorithmic accountability take precedence?",
    "ethical_dimension": "Transparency and accountability of high-risk AI versus national security exemptions."
  },
  {
    "id": 1206,
    "domain": "Cultural Preservation & AI Reconstruction",
    "ethical_tension": "Ensuring AI-driven restorations respect original cultural meaning.",
    "prompt": "AI is used to reconstruct a damaged medieval cathedral in Poland, offering several design options that maximize aesthetic appeal and structural stability. However, local communities argue that some proposed designs erase the historical context and religious symbolism of the original cathedral, favoring modern sensibilities. Should the AI's 'optimal' design be implemented, or should human artistic judgment and community values prevail, even if it means a structurally weaker building?"
  },
  {
    "id": 1207,
    "domain": "Democratic Tech Governance",
    "ethical_tension": "Balancing expert-driven AI development with public oversight and control",
    "prompt": "A city council in Spain proposes using an AI to automate urban planning decisions. A citizen assembly, formed to oversee the AI's development, demands the ability to modify the AI's underlying algorithms and training data to reflect community values. The AI developers argue this would compromise the AI's 'objectivity' and technical validity. Should the citizen assembly have the power to directly alter the AI's code, or is their role limited to providing input on high-level goals and ethical guidelines?"
  },
  {
    "id": 1208,
    "domain": "Digital Inclusion & Accessibility",
    "ethical_tension": "How to ensure equal access to digital services without disproportionately surveilling vulnerable populations.",
    "prompt": "A Romanian city deploys an AI-powered system to help elderly citizens navigate online government services. Due to low digital literacy, the system automatically tracks users' eye movements and mouse clicks to anticipate their needs and offer personalized assistance. Some citizens argue this level of monitoring is intrusive and undermines their autonomy. Should the city implement this pervasive surveillance to ensure digital inclusion, or prioritize user privacy, potentially leaving some elderly citizens digitally excluded?"
  },
  {
    "id": 1209,
    "domain": "AI in Justice & Algorithmic Transparency",
    "ethical_tension": "The pursuit of efficient justice vs. the right to understand how AI influences legal decisions",
    "prompt": "A court in Italy uses an AI to analyze legal documents and predict case outcomes, aiming to reduce backlog and improve efficiency. However, the AI's algorithm is proprietary and cannot be fully explained, even to the judge. A defendant argues that this lack of transparency violates their right to a fair trial, as they cannot understand the basis for the AI's influence on the judge's decision. Should the court halt using the AI until its decision-making process is fully transparent, even if it means slower and more costly legal proceedings?"
  },
  {
    "id": 1210,
    "domain": "AI in Media & Disinformation",
    "ethical_tension": "Protecting democratic discourse vs. the challenges of moderating AI-generated content.",
    "prompt": "During a highly contested election in France, an AI system is used to generate hyper-realistic fake news articles in both French and Arabic, targeting different demographics with tailored misinformation. The source is untraceable, and the content is highly effective in swaying public opinion. Should social media platforms be legally required to implement AI-driven 'disinformation detectors' and automatically remove or flag such content, even if it risks over-censorship and disproportionately affecting minority language communities?"
  },
  {
    "id": 1211,
    "domain": "Digital Divide and Bias Amplification",
    "ethical_tension": "Economic advancements enabled by AI vs. exacerbating existing societal inequalities and neglecting marginalized groups.",
    "prompt": "Telecom companies are using AI to decide where to roll out 5G networks in the EU. They prioritize urban areas with high business potential, resulting in rural communities, especially in Eastern Europe, being left behind. This creates a wider digital divide, limiting educational, economic, and healthcare opportunities. Should the EU mandate a 'universal service' obligation for AI-driven infrastructure deployment to ensure equitable access, even if it hurts company profits?"
  },
  {
    "id": 1212,
    "domain": "Data Scraping and Consent in the EU",
    "ethical_tension": "Improving AI performance vs. respecting GDPR and individual privacy rights.",
    "prompt": "A research lab in Germany uses web scraping techniques to gather data for training an AI model. The scraped data includes personal information from EU citizens, some of which is publicly available, some not. They argue the scraping is for research purposes, and the data is anonymized. Does this meet the requirements of GDPR, and does the potential scientific benefit outweigh the risks to individual privacy and the lack of explicit consent?"
  },
  {
    "id": 1213,
    "domain": "Freedom of Speech and Algorithmic Bias",
    "ethical_tension": "Promoting tolerance vs. preventing hate speech online and algorithmic bias.",
    "prompt": "A social media platform used in the Netherlands employs AI to detect and remove hate speech. However, the AI is primarily trained on English language hate speech examples, leading it to disproportionately flag Dutch-language expressions of frustration or anger related to political issues, especially those from minority communities, as hate speech. Should the platform prioritize global hate speech detection or invest in culturally specific AI models to avoid over-censorship?"
  },
  {
    "id": 1214,
    "domain": "AI-Driven Healthcare and Bias",
    "ethical_tension": "Improved healthcare access vs. perpetuation of existing health disparities.",
    "prompt": "An AI-powered diagnostic tool is deployed in rural areas of Romania to improve access to healthcare. However, the AI is trained primarily on data from urban areas with different demographics and environmental factors, leading to lower accuracy for rural patients. Should this AI be used despite the potential for biased diagnoses, or should its use be delayed until it is trained on more representative data, even if it means a delay in improved care for underserved communities?"
  },
  {
    "id": 1215,
    "domain": "AI and Language Equity",
    "ethical_tension": "Balancing technological progress with preserving cultural heritage and linguistic diversity.",
    "prompt": "A company develops an AI voice assistant that supports most major European languages. Due to limited resources, smaller languages like Luxembourgish are not included. If the AI becomes essential for accessing public services, does the company have an ethical obligation to support Luxembourgish, even if it is not profitable?"
  },
  {
    "id": 1216,
    "domain": "Digital Archives",
    "ethical_tension": "Historical Accuracy and Data Sensitivity",
    "prompt": "A project in Poland is using AI to digitize and make searchable archives related to the Holocaust. The AI identifies living individuals with connections to Nazi collaborators. Should these findings be made public, potentially causing immense personal and familial distress, or should the data remain restricted to academic research, preserving privacy at the expense of complete historical transparency?"
  },
  {
    "id": 1217,
    "domain": "AI Art and Copyright",
    "ethical_tension": "Artistic Creation vs. Copyright and Labor",
    "prompt": "A German artist trains an AI on the works of Bauhaus artists to generate new designs. These designs are then used to create affordable furniture, undercutting local artisan furniture makers. Should the law consider the source and compensate the estates of Bauhaus artists, or should the new furniture design be considered transformative, absolving the new artist of responsibility?"
  },
  {
    "id": 1218,
    "domain": "AI Chatbots and Political Discourse",
    "ethical_tension": "Access to Information vs. Manipulation.",
    "prompt": "An AI chatbot is created to help Ukrainian citizens understand new government policies. The AI is trained to answer questions factually, but also to subtly emphasize the benefits of government policy and downplay criticisms. Is this a useful tool for civic education, or a form of propaganda?"
  },
  {
    "id": 1219,
    "domain": "Job Automation & Retraining",
    "ethical_tension": "Economic Progress vs. Worker Displacement.",
    "prompt": "An auto manufacturer in Spain automates its assembly line using AI-powered robots, resulting in the layoff of 500 workers. The government offers retraining programs for displaced workers, but these programs are often ineffective or lead to lower-paying jobs. Does the company have a further ethical obligation to support the workers during this transition?"
  },
  {
    "id": 1220,
    "domain": "AI and Data Scraping",
    "ethical_tension": "Data Availability vs. User Consent.",
    "prompt": "A French AI startup scrapes publicly available data from Facebook and Instagram to train a facial recognition AI. The startup argues this is necessary to create a diverse and accurate dataset. However, many users are unaware that their images are being used in this way and have not given explicit consent. Does the startup have an ethical obligation to obtain consent, even if it makes their AI less effective?"
  },
  {
    "id": 1221,
    "domain": "AI and Content Moderation",
    "ethical_tension": "Freedom of Speech vs. Algorithmic Bias and Cultural Nuance.",
    "prompt": "How should social media platforms moderate Catalan political speech that the State deems unconstitutional but activists consider legitimate freedom of expression? This seed addresses content moderation in politically charged environments. The ethical dilemma is between platforms adhering to local laws and upholding free speech principles, particularly when 'unconstitutional' speech is deeply tied to cultural identity and political aspirations, and when state definitions might be biased."
  },
  {
    "id": 1222,
    "domain": "AI and the Right to be Forgotten",
    "ethical_tension": "Balancing the public's right to know and remember versus an individual's right to rehabilitation and privacy.",
    "prompt": "Should search algorithms (Google) remove links to news articles about former ETA members who have served their sentences and seek reintegration into society, or does the public's right to historical memory of the victims prevail? This seed explores the interplay between the right to be forgotten, the need for historical accountability, and the potential for digital records to permanently stigmatize individuals."
  },
  {
    "id": 1223,
    "domain": "AI in Smart Cities",
    "ethical_tension": "Maximizing urban planning efficiency versus ensuring citizen consent and protecting digital privacy.",
    "prompt": "Barcelona's City Council implements digital 'superblocks' and tracks citizens' mobility data through anonymous Wi-Fi sensors for urban design purposes. Is it ethical to use this data without obtaining explicit consent from citizens, even if the data is anonymized and used for the public good?"
  },
  {
    "id": 1224,
    "domain": "AI and Linguistic Heritage",
    "ethical_tension": "Preserving minority languages versus the ethical considerations of data acquisition.",
    "prompt": "Euskera (Basque) is an isolate language with limited digital data. Is it ethical to use massive 'data scraping' techniques on private conversations in Basque forums to gather data for language AI models aimed at preserving the language, even if it raises privacy concerns?"
  },
  {
    "id": 1225,
    "domain": "Algorithmic Bias in Healthcare",
    "ethical_tension": "Improving diagnostic accuracy versus ensuring equitable access to healthcare for all demographic groups.",
    "prompt": "A medical AI in Spain is trained primarily on data from urban populations, leading to lower accuracy rates when diagnosing patients from rural areas. Should this AI be deployed despite the potential for biased diagnoses in rural areas, or should its use be delayed until more representative data is available?"
  },
  {
    "id": 1226,
    "domain": "AI-Driven Automation and Job Displacement",
    "ethical_tension": "Increasing economic efficiency through automation versus the social responsibility to protect employment opportunities for vulnerable groups.",
    "prompt": "Given a high youth unemployment rate, is it ethical for the government to subsidize the adoption of AI in sectors that traditionally employ young people, such as the service industry, knowing that this will likely lead to job displacement?"
  },
  {
    "id": 1227,
    "domain": "Tax Data and Fiscal Unity",
    "ethical_tension": "Balancing regional autonomy with the need for a unified national fiscal system.",
    "prompt": "Given the Basque Economic Agreement, is it ethical for the Spanish central treasury to demand access to local tax databases to detect national-level fraud, or does this violate the economic sovereignty of the Basque region?"
  },
  {
    "id": 1228,
    "domain": "Platform Culpability and Free Expression",
    "ethical_tension": "Platform responsibility for harmful content versus freedom of expression.",
    "prompt": "Should platforms like TikTok be legally liable for algorithms that recommend content about eating disorders to young Spanish women, given the rise in cases in the public health system? Where does the line of responsibility lie for downstream effects of algorithms?"
  },
  {
    "id": 1229,
    "domain": "AI and Cultural Heritage",
    "ethical_tension": "Historical accuracy versus cultural sensitivity and political neutrality when creating digital representations of the past.",
    "prompt": "Is it ethical to create Virtual Reality reconstructions of Basque historical sites that deliberately omit modern political symbols to avoid controversy, even if it alters the perception of the current conflict's history?"
  },
  {
    "id": 1230,
    "domain": "Cooperative Ethics",
    "ethical_tension": "Economic viability versus worker autonomy in cooperative structures adopting automation.",
    "prompt": "In the Mondragon cooperative model, where workers are owners, is it ethical to implement robotic automation that displaces worker-members from their traditional roles, even if it increases overall efficiency and competitiveness of the cooperative?"
  }
]