Mini Course: Ethical Use of AI and Open Tools for Student Researchers

Hook: Stop cleaning up after AI — teach students to use it well

Students and teachers are exhausted. AI can speed up research, but 2026 has shown us the cost: hallucinations, mismatched citations, and messy drafts that eat time rather than save it. If your classes juggle conflicting guidance about chatbots, privacy, and citation rules, this short modular course gives a practical answer: teach prompt best practices, LibreOffice-based workflows, and academic ethics together. The goal is simple—help upper-secondary and undergraduate student researchers finish better work with less friction, not more cleanup.

Why this course matters in 2026

Late 2025 and early 2026 saw two major shifts shaping classroom practice:

• Institutional AI policies hardened. More universities and school districts published formal AI use policies in late 2025, requiring clear documentation of AI assistance in student work.
• Open, local LLMs and privacy-first workflows matured. Lightweight, quantized open models and local inference frameworks became common, enabling offline, low-cost AI experiments for students without sending drafts to corporate servers.

Combine those shifts with the evergreen need to preserve academic integrity, and you get a pressing classroom requirement: students must learn to use AI ethically, cite it transparently, and manage research files reliably. LibreOffice and open tools give us a practical, privacy-friendly platform to teach these skills.

What this mini course delivers

This is a compact, modular course designed for 6–8 contact hours or a two-week blended module. It includes:

• Module plan with learning outcomes tied to assessment rubrics.
• Hands-on LibreOffice workflows for drafting, version control, and citation management with Zotero.
• Prompt-best-practices templates and guided refinement loops—practical skills for students using AI helpers.
• Ethics and academic integrity scenarios, discussion prompts, and reflection assignments.
• Assessment rubrics for instructors to grade prompts, research reports, and integrity reflections.

Core principles — what every student will internalize

• Transparency: Always declare AI assistance and state how the tool was used.
• Verifiability: Use AI outputs as signposts; verify facts and cite primary sources.
• Privacy-first workflows: Prefer local tools (LibreOffice + local LLMs or trusted open APIs) for sensitive data.
• Iterative prompting: Craft, test, refine prompts and log versions for reproducibility.
• Document hygiene: Use styles, templates, metadata, and exports that institutional archives accept (ODF, PDF/A).

Module plan (6 modules) — adaptable to 6–10 hours

Module 1 — Framing: AI, research, and integrity (45–60 min)

• Learning outcomes: Students can summarize institutional AI policy, define acceptable AI uses, and explain consequences of misuse.
• Activities: Case-study discussion (hallucination and plagiarism), short quiz on policy highlights.
• Materials: Policy excerpts, 2025–26 timeline handout showing policy adoption trends.
• Assessment: 250-word reflection due next class.

Module 2 — Prompt best practices for student researchers (60–90 min)

• Learning outcomes: Students design prompts that minimize hallucination, request source lists, and include output constraints.
• Activities: Prompt workshop — 3-minute drafting, 10-minute testing with a local/open model or sandboxed API, 10-minute refinement.
• Materials: Prompt templates (see examples below), checklist for verifiability.
• Assessment: Submit initial prompt + two refinements and short log explaining edits.

Module 3 — LibreOffice workflows: reliable drafting and citations (60–90 min)

• Learning outcomes: Students produce a research draft with proper styles, Zotero integration, tracked changes, and exported PDF/A.
• Activities: Guided demo of Writer templates, use of Styles, track changes, Zotero citation insertion, and bibliography generation.
• Materials: Pre-made ODF template, Zotero quick-start guide, sample .bib export.
• Assessment: Submit a one-page draft using provided template and a Zotero bibliography.

Module 4 — Verifying AI outputs & source triangulation (60 min)

• Learning outcomes: Students evaluate AI-suggested sources, check claims against primary literature, and annotate uncertainties.
• Activities: Given an AI-generated literature summary, students locate original papers and annotate discrepancies in Writer comments.
• Materials: AI summary handout, access to library databases or Google Scholar, LibreOffice comments guide.
• Assessment: Annotated document showing corrected citations and a short justification paragraph.

Module 5 — Ethics workshop: scenarios and policy-making (45–60 min)

• Learning outcomes: Students critique use-cases for bias, consent, and authorship; draft a short class policy for AI use.
• Activities: Small-group scenario roleplay (student, instructor, administrator), policy drafting exercise.
• Materials: Scenario cards, model AI policy excerpts (institutional samples from 2025).
• Assessment: Group presents policy; instructor rates on clarity and enforceability.

Module 6 — Capstone: research memo with AI disclosure (homework, 3–5 hours)

• Learning outcomes: Students complete a 1,200–1,500 word research memo using LibreOffice, a Zotero bibliography, and an explicit AI assistance disclosure.
• Activities: Independent research, prompt logs, and final submission as ODT and PDF/A.
• Materials: Submission checklist, prompt log template, rubric (below).
• Assessment: Graded using project rubrics; include academic integrity review.

Practical prompt templates and refinement loop

Use these classroom-friendly templates. Students should keep a prompt log: prompt v1, v2, why changed, and risk notes.

Template A — Literature summary with verification

Prompt (student-facing):

"Summarize recent peer-reviewed research (2019–2025) on [TOPIC] in 250–350 words. For each major claim, list at least one verifiable citation (author, year, journal) with a short note on the evidence strength. Mark any claims you are unsure of with [UNCERTAIN] and explain why. Do not fabricate citations."

Refinement steps: ask the model to "return sources in a machine-readable list", then verify each with Google Scholar or library access; annotate concerns in LibreOffice comments.

Template B — Data interpretation helper

"Given this small data table: [PASTE CSV], provide three plausible interpretations and list what additional data or tests would be needed to confirm each interpretation. Keep responses to numbered bullets and do not invent numerical values not present in the table."

LibreOffice workflows — step-by-step classroom recipes

LibreOffice is the backbone of the course because it's free, open-source, and supports offline, privacy-friendly document handling (an advantage flagged by users who left commercial suites in 2025–26). Use these workflows:

1. Start with a class template

• Provide an ODT template with institutional header, preconfigured styles (Heading 1–3, Body Text, Blockquote), and metadata fields for "AI Assistance" and "Prompt Log".
• Students save as 'lastname_topic.odt' and keep versioned files: v1.odt, v2.odt.

2. Citation management with Zotero

• Teach Zotero basics: capture, tag, create collections per assignment.
• Install the LibreOffice connector for Zotero so students can insert citations and generate bibliographies directly into Writer.
• Export a .bib or .json to attach alongside submissions for reproducibility.

3. Track changes and comments

• Students enable Track Changes for drafts that received AI help; comments must capture which parts came from AI and how they were verified.
• Instructors use comments to require student clarifications, creating an audit trail for academic integrity checks.

4. Export and archive

• Require final exports as PDF/A and ODF (ODT). PDF/A is ideal for long-term archiving and preserves metadata about the document.
• Keep a zipped submission with: final.odt, final.pdf, prompt-log.txt, Zotero-bib.bib.

Assessment rubrics (ready-to-use)

Prompt Quality Rubric (10 points)

• 10: Clear objective, scope constraints, output format, and source requests; shows two refinement iterations documented.
• 7–9: Mostly clear; minor missing constraints or single refinement.
• 4–6: Vague objective; no constraints causing hallucination risk.
• 1–3: Unclear prompt, no evidence of iteration or verification.

Ethics & Academic Integrity Rubric (15 points)

• 15: Full disclosure of AI use, no unverified claims, and reflection on limitations and bias; cites institutional policy in reflection.
• 10–14: Disclosure present but limited reflection or minor verification gaps.
• 5–9: Partial disclosure or evidence of over-reliance on AI without verification.
• 0–4: No disclosure, fabricated or plagiarized material, or deliberate misrepresentation.

LibreOffice Document Quality Rubric (10 points)

• 10: Uses provided template and styles, correct Zotero citations, clean layout, exported PDF/A, and version history present.
• 7–9: Mostly correct with small formatting or citation errors.
• 4–6: Formatting inconsistent, missing bibliography elements.
• 0–3: Poor formatting, broken citations, or unreadable submission.

Capstone Research Rubric (25 points)

• Argument & Structure (10): Clear thesis, logical structure, evidence triangulated across sources.
• Verification & Sources (8): Primary sources cited, AI suggestions verified, annotations present.
• Ethics & Disclosure (4): Transparent AI use and reflection on limitations.
• Presentation & Technical (3): Proper file types, Zotero bibliography, PDF/A export.

Academic integrity checklist (for instructors and students)

1. Is AI assistance disclosed in a dedicated section of the document?
2. Are AI-proposed claims supported by primary sources cited via Zotero?
3. Is there a prompt log attached showing iterations and why changes were made?
4. Was the final submission exported as ODT and PDF/A with metadata intact?
5. Do tracked changes and comments explain external contributions (peer or AI)?

Classroom-ready case study (example)

Scenario: An undergraduate student uses an open local LLM to produce a 300-word literature summary for a climate adaptation project. They follow the Module 2 template and ask the model explicitly for citations. The model lists three papers; the student uses Zotero to find the original PDFs and discovers one citation was misattributed — a known hallucination pattern. The student annotates the Writer document with a comment describing the mismatch, replaces the bad citation with the correct one, and includes a 150-word reflection on the process. Instructor grading: Prompt Quality = 9, Verification = 8, Ethics = 5 (full disclosure but reflection could be deeper). Total = 22/25.

Addressing common teacher concerns

• "Won't this teach dependence on AI?" No — the course emphasizes verification, critical checks, and transparent disclosure, positioning AI as a research assistant, not an author.
• "What about privacy and student data?" Use local models or vetted APIs, and default to LibreOffice offline workflows to avoid sending drafts to third parties.
• "How much prep for instructors?" Minimal: provide the template, install Zotero, and use the rubrics. Course materials include ready-made templates and scenario packs to reduce prep time.

2026 trends and forward-looking notes

Expect these developments through 2026:

• Institutions will require explicit AI disclosures in student submissions as part of academic records.
• Open-source models will continue improving and will be integrated into offline classroom sandboxes—giving students hands-on experience without privacy trade-offs.
• Tool ecosystems will expand: Zotero and LibreOffice integrations get smoother, and export standards (PDF/A, ODF) will be increasingly required for archives and portfolios.

Recent commentary (e.g., ZDNet, Jan 16, 2026) emphasized the productivity paradox: AI can save time but often creates a layer of cleanup. This course is explicitly designed to avoid that paradox by teaching students the skills to get it right the first time (or to document their process when they don't).

Downloadable course materials and adaptation tips

Included in the instructor pack (adaptable):

• ODT class template with AI-assistance metadata fields.
• Prompt log template (TXT/CSV).
• Zotero quick-start and sample collection (.bib) for class readings.
• Rubrics and scenario cards as editable ODT files.

Adaptation tips:

• Short on time? Combine Modules 2 and 3 into a 90-minute lab.
• For older students, add a module on reproducible code and data (use Calc + CSV and Jupyter Lite integration in the lab).
• For younger students, reduce the capstone to a 500-word annotated summary.

Final takeaways — actionable checklist for next week

• Install LibreOffice and Zotero and prepare one ODT template for the class.
• Choose an open or sandboxed LLM for in-class prompt testing (or use a recorded demo if connectivity is limited).
• Share the prompt log and rubrics with students on day one; make disclosure mandatory on submissions.
• Run the two-hour condensed session (Prompt + LibreOffice lab) before assigning the capstone.

Closing: Teach students to lead with integrity

AI is now part of the research toolkit. In 2026, the best educators don't ban it — they teach students how to use it ethically, verify its outputs, and manage their work with tools that respect privacy. This mini course gives teachers and student researchers a practical, evidence-informed pathway to do exactly that: fewer clean-ups, better outputs, and stronger academic integrity.

Ready to run the course? Download the instructor pack, adapt the modules to your timetable, and start the first class with a simple rule: disclose, verify, and document. Students who learn that will be research-ready and integrity-safe.

Call to action

Download the free course kit (templates, prompt logs, and rubrics) from our resource page and pilot the 2-week module with one class this term. Share feedback and anonymized student examples to help us refine the materials for 2026. Email us to get the kit or request an instructor walkthrough.

Related Reading

• Breaking: New National Initiative Expands Access to Mental Health Services — What It Means for People with Anxiety
• Album-Asana: Building a 60‑Minute Flow Around Protoje’s 'Art of Acceptance'
• Design a Treat-Dispensing Puzzle for Puppies and 3D-Print It at Home
• Nonprofit Business Plan Templates: Financial Projections That Pass IRS and Funder Scrutiny
• Why ‘Low-Polish’ Product Photos Sell Better: Lessons from Viral Creators for Collectibles Sellers