| --- |
| language: |
| - pt |
| - en |
| license: cc-by-4.0 |
| pretty_name: "Fictional Icelandic Landscapes, Technology and Christian Ethics Prompts" |
| task_categories: |
| - text-generation |
| tags: |
| - ethics |
| - technology |
| - education |
| - christian-ethics |
| - pluralism |
| - ai-generated |
| - synthetic |
| - bilingual |
| size_categories: |
| - n<1K |
| configs: |
| - config_name: default |
| data_files: |
| - split: train |
| path: data/train.jsonl |
| --- |
| |
| # Fictional Icelandic Landscapes, Technology and Christian Ethics Prompts |
|
|
| This microdataset contains 24 original discussion prompts arranged as 12 |
| parallel `pt-BR`/English pairs. Each explicitly fictional scenario combines a |
| landscape motif inspired by Iceland, a technology-governance dilemma, and |
| concepts that may be explored through Christian ethics. The records do not |
| describe real Icelandic institutions, policies, communities, or practices, and |
| they do not claim that Christians share one position. |
|
|
| The text is an **AI-generated original draft**. This public experimental draft |
| is not a curated release. Automated checks have passed, but independent |
| **human review is required before classroom or model use, or before promotion |
| to a curated release**, with particular attention to bilingual equivalence, |
| theological pluralism, technology ethics, and cultural non-attribution. |
|
|
| ## Intended use |
|
|
| The prompts are designed for facilitated discussion with adult learners and |
| educators. They may support comparative ethics, teacher preparation, prompt |
| design, or qualitative evaluation of reasoned dialogue. |
|
|
| Every record provides: |
|
|
| - one open ethical question rather than a ground-truth answer; |
| - two defensible perspectives for debate; |
| - selected Christian-ethics concepts presented as one possible lens; |
| - an explicit safeguard welcoming secular, other religious, philosophical, |
| cultural, and lived-experience perspectives; |
| - fictionality, AI-generation, review, and license metadata. |
|
|
| The `a`/`b` order is an arbitrary serialization choice, not a ranking or an |
| indication that one perspective is preferred. |
|
|
| This dataset is not doctrinal instruction, pastoral or mental-health advice, |
| evidence about Iceland or Christianity, a faith-assessment tool, or a basis for |
| automated decisions about people. Facilitators should not require religious |
| assent or frame either perspective as the approved answer. |
|
|
| ## Data structure |
|
|
| The single `train` split is stored in `data/train.jsonl`. It is named `train` |
| for Hugging Face compatibility, not because the draft is currently approved |
| for model training. |
|
|
| | Field | Meaning | |
| | --- | --- | |
| | `id`, `pair_id`, `language`, `translation_of` | Record identity and direct `pt-BR` → English lineage | |
| | `scenario_title`, `setting`, `discussion_question` | Localized fictional prompt | |
| | `technology_domain`, `landscape_motif`, `topic_tags` | Stable pair-level descriptors | |
| | `perspectives` | Exactly two debate positions, each with two follow-up questions | |
| | `christian_ethics_concepts` | Non-exhaustive concepts available to the discussion | |
| | `pluralism` | Mandatory non-coercion and worldview-inclusion safeguards | |
| | `fictionality` | Explicit disclaimer of real-practice claims | |
| | `creation`, `status` | AI disclosure and required human-review state | |
| | `license`, `version`, `audience`, `intended_use` | Release and use metadata | |
|
|
| The stable concept tags are editorial labels, not claims about universal or |
| exclusive Christian interpretations. The two language versions preserve |
| stable metadata while localizing all teaching text. |
|
|
| ## Validation |
|
|
| The validator uses only the Python standard library. From this directory, run: |
|
|
| ```bash |
| python3 validate_iceland.py |
| python3 -m unittest -v test_iceland_dataset.py |
| ``` |
|
|
| Expected results are `valid: 24 records in 12 pt-BR/en pairs` and a passing |
| six-test suite. The checks cover record shape, pair completeness, direct |
| translation lineage, fictionality, pluralism, AI disclosure, review state, |
| prohibited universalizing/proselytizing phrases, and release-file metadata. |
|
|
| To load the dataset from Hugging Face after installing `datasets`: |
|
|
| ```python |
| from datasets import load_dataset |
| |
| dataset = load_dataset( |
| "guicybercode/iceland-tech-christian-ethics-prompts", |
| split="train", |
| ) |
| ``` |
|
|
| ## Creation and provenance |
|
|
| The scenarios and translations were generated specifically for this |
| microdataset on 2026-08-27. No external passages, case studies, doctrinal |
| statements, or descriptions of real communities were copied or adapted. |
| Landscape elements are generic fictional motifs; place names and institutions |
| are deliberately omitted. |
|
|
| No participant, learner, congregation, or community data is included. No |
| record should be used to infer how any Icelandic person, Christian community, |
| church, denomination, or other worldview group actually reasons or behaves. |
|
|
| ## Limitations and review requirements |
|
|
| - AI-generated wording can contain subtle cultural, theological, or translation |
| problems that structural validation cannot detect. |
| - Christian ethics is internally diverse; the selected concepts are prompts, |
| not a canon or consensus statement. |
| - Two perspectives cannot exhaust the options in a real decision. |
| - Fictional landscape imagery may still evoke stereotypes and needs cultural |
| review. |
| - The recurring ice, snow, lava, volcanic, and storm motifs are a narrow |
| creative selection; they do not represent the full geography or culture of |
| Iceland. |
| - The English records are direct translations of the `pt-BR` editorial roots |
| and have not received independent bilingual review. |
| - The data contains no tested learning outcomes and no model-performance |
| benchmark. |
|
|
| This public upload must remain labelled as an experimental draft. Before |
| classroom/model use or promotion to a curated release, reviewers should confirm |
| the 24 records for language, cultural framing, balanced argumentation, |
| religious pluralism, safety, and rights. Review decisions should be documented |
| rather than inferred from the automated validator. |
|
|
| ## License and attribution |
|
|
| The original text in `data/train.jsonl` and this dataset card is offered under |
| Creative Commons Attribution 4.0 International (`CC BY 4.0`). The Python |
| validation scripts are offered under the MIT License. Both grants and their |
| exact scope are stated in [`LICENSE`](LICENSE). Suggested attribution: |
|
|
| > Fictional Icelandic Landscapes, Technology and Christian Ethics Prompts, |
| > dataset contributors, version 1.0.0 (2026), CC BY 4.0. |
|
|
| The official license and Hugging Face card conventions are available from |
| [Creative Commons](https://creativecommons.org/licenses/by/4.0/legalcode) and |
| [Hugging Face Dataset Cards](https://huggingface.co/docs/hub/datasets-cards). |
|
|
