Research-integrity, Ethics, and Source-Use Reminder
Table of Contents
- 1. Lesson Overview
- 2. Learning Outcomes
- 3. Before You Begin
- 4. Retrieval Practice
- 5. Opening Scenario
- 6. Core Concepts
- 7. RIGOUR Walkthrough
- 8. Research-Process Walkthrough
- 9. Worked Example
- 10. Guided Learning Activities
- 11. Exercises
- 12. Homework Problems
- 13. Mini-Project
- 14. Self-Check Quiz
- 15. Assessment Connection
- 16. Assignment 1 Journal Connection
- 17. Key Takeaways
- 18. Key Terms
- 19. Looking Ahead
- 20. Sources and Further Reading
1. Lesson Overview
This lesson moves from research planning to the systematic discovery, verification, and synthesis of scholarly literature. You will learn how generative AI may assist with search-term development, literature discovery, and organisational mapping—but you will also learn why verification is a non-negotiable human responsibility. AI can suggest sources, but only you can confirm that they exist, are accurately represented, and are relevant to your research question.
You will distinguish between a literature list (a simple inventory of sources), a literature map (an organised thematic structure), and a scholarly synthesis (an integrated, evidence-based account of patterns, disagreements, and gaps). You will also learn to detect fabricated sources, citation laundering, and selective coverage—practices that undermine research integrity.
Literature discovery without verification is not scholarship—it is speculation. A credible researcher does not simply find sources; they verify, situate, and synthesise them in a way that respects evidence, disagreement, and context. This lesson equips you with the skills to use AI as a discovery assistant while maintaining rigorous human oversight.
2. Lesson Learning Outcomes
By the end of this lesson, you will be able to:
- Use AI cautiously to develop candidate search terms and controlled vocabulary.
- Distinguish literature discovery from source verification.
- Verify that every cited source exists and is represented accurately.
- Distinguish a source list, annotated bibliography, literature map, summary, and scholarly synthesis.
- Construct a literature map showing themes, methods, findings, disagreements, limitations, and gaps.
- Identify selective coverage, confirmation bias, citation laundering, and fabricated literature.
- Synthesise literature without erasing disagreement, context, methods, or limitations.
- Apply TRACE and CLAIM to verify literature claims and citations.
3. Before You Begin
- Assigned reading: Read the literature review sections of Chapter 14, including the subsections on literature mapping and synthesis. Review Chapter 7 (TRACE) and Chapter 8 (CLAIM) for verification protocols.
- Prerequisite knowledge: You should be able to formulate a focused research question (Lesson 13.1) and apply TRACE and CLAIM to identify errors and verify claims (Units 10–11).
- Library and public-search options: You may use your institution's library, open-access repositories (e.g., Google Scholar, arXiv, SSRN, PubMed Central), or approved AI tools if they are available. No activity requires a paid database or a specific AI system.
- Optional AI tool: If you have access to an approved generative AI system, you may use it for search-term brainstorming and discovery. Do not treat any AI-generated reference as verified.
- Non-platform alternative: Every activity includes a complete supplied source set and synthetic data so that you are not disadvantaged if you choose not to use an AI system or paid database.
- Privacy and confidentiality: Use only synthetic, public, supplied, or explicitly authorised information.
- Source-use reminder: Respect intellectual property and community-governed knowledge.
- Accessibility: All diagrams have text descriptions.
- Preparation time: Approximately 15–20 minutes.
4. Retrieval Practice
These questions activate knowledge from Lesson 13.1 and earlier units. Attempt each one before revealing the feedback.
Question 1: In Lesson 13.1, you learned to distinguish a research topic from a research problem. Why is this distinction important for literature discovery?
Show answer
Answer: A topic is too broad to guide a focused literature search. A research problem narrows the scope to specific concepts, populations, and contexts, which enables the development of precise search terms and inclusion criteria.
Question 2: Using TRACE (Unit 10), what should you do when you encounter an AI-generated reference list that seems too comprehensive?
Show answer
Answer: Apply TRACE by tracing each reference: verify that the source exists, the citation is accurate, and the content actually supports the claim.
Question 3: From Unit 11 (CLAIM), what is the first step in verifying a claim made in a research article?
Show answer
Answer: The first step is to locate the original source of the claim and verify that the claim is accurately represented in context.
Discovery is about finding potential sources; verification is about confirming that those sources are real, accurate, relevant, and appropriately represented. Discovery without verification is a risk, not a foundation.
5. Opening Scenario
Scenario: Jamil is a graduate student in public policy. He has a research question: "How do municipal governments in Canada use generative AI for public engagement?" To accelerate his literature review, he asks an AI system to "provide a comprehensive literature review on AI and public engagement in Canadian municipalities." The AI returns:
- A polished list of 25 references with authors, titles, journals, and DOIs.
- One fabricated article that looks real but does not exist.
- Several sources that are real but irrelevant (e.g., a study on AI in US federal agencies).
- Three sources derived from the same systematic review, giving an illusion of breadth.
- A claim that "the literature agrees" that AI improves public engagement, despite conflicting evidence in the field.
Task: Identify all the problems in this AI-generated literature review. What should Jamil do to salvage a credible review?
Show guided analysis
Problems identified:
- Fabricated source: The AI invented an article.
- Irrelevant sources: Sources on US federal agencies do not answer a question about Canadian municipalities.
- Source dependence: Citing three sources from the same review creates a false impression of breadth.
- False consensus: The claim that "the literature agrees" obscures real disagreements.
- No verification: The AI did not provide any verification metadata.
What Jamil should do:
- Use the AI-generated list as a starting point for discovery, not as a final product.
- Verify each reference by checking library catalogs, Google Scholar, and journal websites.
- Read the abstracts and full texts to assess relevance.
- Identify the main themes, methods, findings, and disagreements.
- Construct his own literature map and synthesis.
6. Core Concepts
6.1 Search-Term Development
A search strategy is a systematic plan for identifying relevant literature. It begins with key concepts from your research question and expands to include:
- Synonyms – alternative terms for the same concept (e.g., "generative AI" → "large language models," "LLMs," "foundation models").
- Controlled vocabulary – standardised terms used in databases (e.g., MeSH terms in medicine, ERIC descriptors in education).
- Related concepts – broader, narrower, or related ideas.
- Disciplinary variations – terms may differ across fields.
- Spelling and language variants – "organisation" vs. "organization."
AI may help generate candidate search terms, but you must evaluate and refine them. Do not assume that the AI's list is comprehensive or appropriate for your discipline.
6.2 Discovery vs. Verification
| Literature Discovery | Source Verification |
|---|---|
| Identifying potential sources (e.g., via search engines, AI suggestions, citation chaining). | Confirming that a source exists, the citation is accurate, and the content is represented correctly. |
| AI can assist by generating candidate references, search terms, and broad topic maps. | AI cannot be trusted to verify sources. Verification requires human access to the original. |
| Discovery is an exploratory activity. | Verification is a confirmatory activity. |
| Discovery may include unverified, irrelevant, or fabricated items. | Verification filters out fabrication, misrepresentation, and irrelevance. |
Critical principle: Discovery is not verification. An AI-generated reference list is a starting point, not a source of truth.
6.3 Literature List, Map, Summary, and Synthesis
| Term | Definition | Depth |
|---|---|---|
| Literature list | A simple inventory of sources. | Low – no analysis or organisation. |
| Annotated bibliography | A list with brief summaries or evaluations of each source. | Low-moderate – individual source description. |
| Literature map | A visual or thematic organisation of sources by theme, method, population, finding, disagreement, limitation, or gap. | Moderate – sources are grouped and related. |
| Scholarly summary | A condensed account of a single source, preserving purpose, method, findings, scope, limitations, and uncertainty. | Moderate – faithful abstraction of one source. |
| Scholarly synthesis | An integrated, evidence-based account of patterns, differences, debates, methods, strengths, limitations, gaps, and implications across multiple sources. | High – critical integration across the field. |
Key distinction: A synthesis is not a series of disconnected summaries. It connects sources, identifies patterns and disagreements, and builds a coherent narrative.
6.4 Literature Map Structure
A literature map organises sources according to:
- Theme – What topics or concepts are addressed?
- Theory – What theoretical frameworks are used?
- Method – What research methods were employed?
- Population – Who or what was studied?
- Context – Where and when was the study conducted?
- Period – Does the literature show changes over time?
- Finding – What were the main results?
- Disagreement – Where do sources conflict?
- Limitation – What are the weaknesses or gaps in the research?
- Gap – What remains unknown or under-studied?
A map supports analysis but does not replace reading the original sources.
6.5 Selective Coverage and Confirmation Bias
Selective coverage occurs when a literature review includes only sources that support a preferred narrative, omitting contradictory or methodologically different studies. AI-assisted discovery may inadvertently privilege:
- Highly visible sources (e.g., high-impact journals).
- English-language sources.
- Frequently cited sources.
- Dominant disciplinary perspectives.
- Sources represented in the AI's training data or search index.
Confirmation bias occurs when researchers consciously or unconsciously seek evidence that supports their existing beliefs. Mitigation: actively seek contradictory evidence, include sources that challenge assumptions, and explicitly address disagreements.
6.6 Citation Laundering
Citation laundering refers to the practice of citing a source without having read it, often by copying citations from other reviews. This can perpetuate errors, misrepresentations, and fabricated sources. AI may amplify citation laundering by generating references that are not verified.
- References that are difficult to locate.
- Citations that seem too perfect or too general.
- Sources that are consistently cited in the same way across multiple reviews.
- References where the cited content does not match the original.
6.7 Applying TRACE and CLAIM to Literature
TRACE (from Unit 10) helps identify errors in AI-generated literature:
- T – Trace the claim back to the original source.
- R – Recognise hallucinations, fabrications, and omissions.
- A – Assess the validity of the reasoning and evidence.
- C – Check for consistency across sources.
- E – Evaluate uncertainty and limitations.
CLAIM (from Unit 11) provides a verification protocol:
- C – Confirm the claim is supported by the cited evidence.
- L – Locate the original source.
- A – Analyse the source's authority, relevance, and currency.
- I – Inspect for misrepresentation or selective quoting.
- M – Map the claim to the evidence and methodology.
For every AI-generated reference or literature claim:
- Trace the claim to the original source.
- Verify that the source exists and the citation is accurate.
- Read the abstract or relevant section to confirm representation.
- Evaluate the source's authority, relevance, and currency.
- Contextualise the source within the broader literature.
- Document your verification.
6.8 Coverage and Omission Audit
A coverage audit examines what is included and what is missing. Consider:
- Dominant perspectives – Are some viewpoints over-represented?
- Absent methods – Are certain research methods excluded?
- Missing populations – Are specific groups or communities omitted?
- Language limitations – Are non-English sources underrepresented?
- Geographic concentration – Is the literature from a narrow set of countries?
- Disciplinary gaps – Is the literature dominated by one discipline?
- Inaccessible or unpublished evidence – Is important work not retrievable?
- Community-governed knowledge – Are Indigenous or local knowledge systems excluded?
Documenting these omissions is part of scholarly integrity.
7. RIGOUR Walkthrough
In Lesson 13.2, we focus on the Information, data, source, and ethics governance element of RIGOUR, with particular attention to literature discovery, verification, and synthesis.
Definition: The researcher must govern all information, data, and sources according to principles of authority, consent, community rights, and intellectual integrity.
Guiding questions:
- Have I verified every source I cite against the original?
- Does my literature coverage reflect the diversity of methods, populations, and perspectives?
- Have I included and represented conflicting evidence fairly?
- Have I respected community-governed knowledge and data?
Interdisciplinary example: A researcher in health policy uses AI to generate a list of studies on telehealth in Indigenous communities. The researcher verifies each source, discovers that three are fabricated, and also finds that the AI omitted studies published in community-based journals. The researcher includes those community-authored sources after obtaining appropriate permission.
Common misuse: Accepting AI-generated references without verification, or selectively including only sources that support a preferred conclusion.
Verification requirement: Every source must be traced to the original publication. Metadata must be confirmed. The claim must be checked against the original.
Escalation condition: If you cannot verify a source, or if it appears fabricated, escalate to your instructor or librarian.
Residual limitation: Even verified sources may be misinterpreted or used out of context.
Show application exercise: RIGOUR source governance
Prompt: You have an AI-generated reference list of 15 sources. Using RIGOUR's "information governance" element, outline a verification plan.
Sample plan:
- For each reference, check library catalogs or Google Scholar to confirm existence.
- Verify that the citation is accurate.
- Read the abstract and, where possible, the full text to confirm relevance.
- Check whether the claim attributed to the source is actually present.
- Document your verification in a research log.
- Remove any fabricated, irrelevant, or misrepresented sources.
- Note any gaps in coverage.
8. Research-Process Walkthrough: From Question to Verified Literature Map
This walkthrough traces the full literature discovery, verification, mapping, and synthesis process for a bounded research question.
Stage 1: Research question
Question: "How do undergraduate students in a first-year composition course use generative AI to revise their drafts, and what revisions do they report making?"
Stage 2: Identify key concepts
- Undergraduate students
- First-year composition
- Generative AI
- Revising drafts
- Reported revisions
Stage 3: Develop search terms
Human activity: The researcher brainstorms synonyms and variations.
- Undergraduate students → first-year, college, university, post-secondary
- Generative AI → large language models, LLMs, AI writing tools, ChatGPT
- Revising drafts → editing, revision, rewriting, draft improvement
- Reported revisions → self-report, survey, interview, perceptions
Bounded AI activity: The researcher asks: "What are some synonyms and related terms for 'revision' in writing research?" The AI suggests "editing, rewriting, proofreading, global revision, local revision." The researcher evaluates and selects terms.
Stage 4: Search databases
Human activity: The researcher searches library databases using Boolean operators.
Bounded AI activity: The researcher may use an AI-assisted search tool to expand results, but does not accept AI-generated references without verification.
Stage 5: Assemble candidate sources
The researcher collects 20 candidate sources from the database search.
Stage 6: Verify each source
- Existence check: Confirm each source exists in a library catalog or journal website.
- Metadata verification: Check author, title, year, journal, volume, pages, DOI.
- Relevance check: Read abstracts to confirm relevance.
- Claim check: Verify the claim against the original.
Result: 12 sources remain; 8 are removed (3 fabricated, 3 irrelevant, 2 duplicates).
Stage 7: Build a literature map
The researcher organises the 12 sources by:
- Theme: (1) AI capabilities for revision, (2) student perceptions, (3) pedagogical implications.
- Method: Surveys (6), interviews (4), case studies (2).
- Findings: Most studies report positive perceptions, but some note quality concerns.
- Disagreements: Two studies found that AI reduces critical revision; two others found it enhances it.
- Gaps: No studies focus specifically on first-year composition courses.
Stage 8: Write a qualified synthesis
Research on generative AI for writing revision is emerging and predominantly positive in tone. Survey-based studies report that students perceive AI tools as helpful for editing and error correction. However, qualitative interview studies reveal concerns about over-reliance and loss of critical editing skills. Methodologically, the field is dominated by surveys, with fewer in-depth qualitative or experimental studies. A notable gap exists in research focusing specifically on first-year composition, as most studies target graduate or professional writers. The evidence is mixed on whether AI improves or diminishes the quality of revision, and this question merits further investigation.
The researcher moved from a research question, to search terms, to discovery, to verification, to mapping, to a qualified synthesis. At every stage, human judgment governed the process.
9. Worked Example
This worked example shows how a researcher discovers, verifies, maps, and synthesises literature on a bounded question.
Research question: "How do public health communicators in Canada perceive the usefulness of generative AI for creating public health messages during health emergencies?"
Step 1: Identify key concepts and search terms
- Population: public health communicators, health communication professionals, Canada.
- Intervention: generative AI, large language models, AI writing tools, ChatGPT.
- Outcome: perceptions, usefulness, trust, credibility.
- Context: health emergencies, public health messaging, crisis communication.
Step 2: AI-assisted discovery
The researcher asks an AI system: "Suggest additional search terms for a literature review on AI and public health communication." The AI suggests: "health crisis communication, emergency risk communication, AI-generated health content." The researcher adds these to the search.
Verification: The researcher checks whether these terms are used in the literature.
Step 3: Source verification
The AI system returns a list of 18 references. The researcher:
- Checks each reference in Google Scholar and library catalogs.
- Finds that 2 references are fabricated (no record exists).
- Finds that 3 references are real but focus on US or European contexts.
- Confirms 13 references as real and relevant.
Step 4: Literature map
| Theme | Sources | Methods | Key Findings |
|---|---|---|---|
| AI capabilities for message generation | A, B, C | Experiments, content analysis | AI can generate clear, accurate messages but may omit nuance. |
| Trust and credibility of AI-generated messages | D, E, F | Surveys, focus groups | Public trust is mixed; credibility depends on source attribution. |
| Practitioner perceptions | G, H, I | Interviews, qualitative | Communicators see potential but worry about accuracy and bias. |
| Ethical and governance issues | J, K, L | Policy analysis, case study | Data privacy, bias, and accountability are key concerns. |
| Canadian context | M | Survey | Only one study focuses on Canada; limited evidence. |
Gap identified: Very little research on Canadian public health communicators.
Step 5: Qualified synthesis
The literature on generative AI in public health communication is growing but geographically concentrated. Most studies originate from the US, Europe, or Australia, with only one Canadian study identified. Findings consistently highlight AI's potential to generate rapid, accessible health messages, but concerns about accuracy, bias, and public trust are equally prominent. Practitioner perceptions are cautiously optimistic, with many communicators expressing a desire for guidelines and training. Ethical governance frameworks are under-developed, and there is a notable absence of research on how AI tools interact with culturally and linguistically diverse populations in Canada.
The researcher verified every source, removed fabricated and irrelevant items, organised the remaining sources into a thematic map, identified a gap, and wrote a synthesis that preserves uncertainty and disagreement.
10. Guided Learning Activities
Instructions: For the research question below, identify key concepts, synonyms, controlled vocabulary terms, and related concepts. Use an AI system if you wish, but evaluate its suggestions.
"How do remote workers in Canadian tech companies perceive the impact of generative AI on their productivity?"
Show sample search-term development
Key concepts: remote workers; Canadian tech companies; generative AI; productivity; perceptions.
Synonyms and variations:
- Remote workers → teleworkers, distributed workforce, virtual workers
- Canadian tech companies → technology firms, IT sector, software companies
- Generative AI → large language models, AI writing tools, ChatGPT, foundation models
- Productivity → efficiency, performance, output
- Perceptions → attitudes, beliefs, experiences, views
Controlled vocabulary: "Telecommuting" (ERIC), "Artificial Intelligence" (ERIC/PsycINFO), "Job Performance" (PsycINFO), "Technology Acceptance" (ACM).
Instructions: Classify each activity as discovery, verification, or both. Explain your reasoning.
- Using Google Scholar to find articles on a topic.
- Checking that an article exists in the library catalog.
- Reading an abstract to determine relevance.
- Confirming that a quote attributed to a source appears in the original.
- Asking an AI to suggest 10 relevant studies.
- Cross-checking a DOI against the publisher's website.
Show classifications
- Discovery.
- Verification.
- Discovery.
- Verification.
- Discovery.
- Verification.
Supplied sources (synthetic):
- Study A: Survey of 200 nursing students on AI confidence (2024). Finding: moderate confidence, desire for training.
- Study B: Interview study with 15 nursing educators on AI integration (2023). Finding: educators are skeptical.
- Study C: Experimental study comparing AI-assisted vs. traditional diagnosis training (2024). Finding: AI group performed better.
- Study D: Systematic review of AI in nursing education (2023). Finding: evidence base is weak.
- Study E: Survey of 500 nursing students on AI ethics concerns (2024). Finding: privacy and bias concerns.
Task: Organise these sources into a literature map with at least three themes. Identify disagreements, limitations, and gaps.
Show sample literature map
| Theme | Sources | Methods | Findings |
|---|---|---|---|
| Student perceptions and confidence | A, E | Surveys | Moderate confidence; privacy and bias concerns. |
| Educator perspectives | B | Interviews | Skepticism due to lack of evidence. |
| AI performance | C | Experiment | AI group outperformed traditional group. |
| Evidence synthesis | D | Systematic review | Evidence base is weak. |
Disagreements: Students optimistic (A, E), educators skeptical (B).
Gaps: No studies on clinical practice integration; no Indigenous health contexts.
11. Exercises
Complete these eight exercises. Each includes a solution or solution design inside a details element.
Task: For the research question "How do public librarians perceive the usefulness of generative AI for reference services?", identify at least three key concepts. For each, list 3–5 synonyms or related terms.
Show sample answer
- Public librarians: librarians, information professionals, library staff
- Generative AI: large language models, AI chatbots, ChatGPT
- Reference services: information services, reader advisory, library instruction
Task: A researcher uses an AI system to generate a list of 30 references. The researcher copies these directly into a bibliography. What has the researcher omitted? What risks does this create?
Show analysis
Omitted: Verification of every reference.
Risks: Fabricated sources, irrelevant sources, misrepresentation, confirmation bias.
Task: Verify the following AI-generated reference: "Smith, J. (2023). The role of AI in community health. Journal of Community Health, 48(3), 234-245. DOI: 10.xxxx/community.2023.001." What steps did you take? What did you find?
Show verification process
- Search for the DOI in the DOI resolver.
- Search the title in Google Scholar.
- Check the journal's website.
- Search the author and title in the library catalog.
Typical findings: The DOI does not resolve, and no record exists. The reference is likely fabricated.
Task: Using the supplied source set, construct a literature map with at least three themes. Identify a disagreement and a gap.
Supplied sources:
- Study 1: Survey of 150 teachers on AI grading tools (2024). 70% found AI helpful; 30% concerned about accuracy.
- Study 2: Interview study with 20 teachers (2023). Cautious; emphasise human oversight.
- Study 3: Experimental study (2024). AI matched teacher grades on 85% of assignments.
- Study 4: Systematic review (2023). Evidence is mixed.
Show sample literature map
Disagreement: Surveys and interviews suggest teacher concern about accuracy, but experiment shows high match rate.
Gap: No studies examine the impact on student learning outcomes or equity concerns.
Task: The following text is either a summary or a synthesis. Identify which it is, and explain why.
"Study A found that AI improved diagnostic accuracy. Study B found that AI reduced diagnostic time. Study C found that AI increased clinician confidence. Study D found no significant difference. Overall, the evidence suggests that AI may enhance some aspects of diagnosis, but the effect is inconsistent."
Show analysis
Classification: Synthesis. It integrates multiple sources, identifies patterns, and acknowledges inconsistency.
Task: A researcher cites the same systematic review 15 times. What is the problem, and what should the researcher do?
Show analysis
Problem: Over-reliance on a single source creates a false impression of breadth.
What to do: Trace the primary studies cited in the review and cite them directly.
Task: A researcher's literature review includes only studies showing AI improves student outcomes, excluding studies showing no effect. What is this an example of? What should the researcher do?
Show analysis
This is: Selective coverage and confirmation bias.
What to do: Actively search for contradictory evidence; revise to include studies with mixed or negative findings.
Task: Using the literature map from Exercise 4, write a synthesis statement on AI grading tools. Preserve disagreement, limitations, and uncertainty.
Show sample synthesis
Synthesis: Research on AI-based grading tools is nascent and characterised by mixed evidence. Survey and interview studies indicate that teachers see potential benefits but harbour concerns about accuracy. Experimental evidence suggests AI can match teacher grading on a substantial proportion of assignments, but this finding is based on a single study. A systematic review cautions that the evidence base is weak. Notable gaps include research on student learning outcomes and equity implications.
12. Homework Problems
Show sample answer
- Social workers: social service workers, caseworkers
- Rural Canada: rural, remote, northern, non-urban
- AI: generative AI, large language models, ChatGPT
- Case documentation: case notes, record-keeping, clinical notes
Show analysis
Sufficiency: The AI's list is broad but lacks specificity. Additional terms: "clinical AI ethics," "AI bias in healthcare," "patient consent AI," "algorithmic fairness medicine," "healthcare AI governance."
Show analysis
- Fabricated sources may be cited.
- Misrepresented sources may lead to incorrect conclusions.
- Academic misconduct charges.
- Irreproducible research.
- Confirmation bias reinforced.
Show verification plan
- Confirm the journal exists.
- Search for the article title.
- Search for the author.
- Check the DOI.
- Confirm page numbers and abstract.
- Flag as potentially fabricated if not found.
- Study A: Survey of 200 HR professionals on AI in recruitment. 65% use AI; concerns about bias.
- Study B: Interview study with 25 HR professionals. Mixed views.
- Study C: Experimental study. AI had 12% lower interview rate for minority candidates.
- Study D: Systematic review. Strong evidence of bias.
Show sample map
Themes: Adoption and perceptions (A, B); Bias and fairness (C); Evidence synthesis (D).
Gap: Few studies test interventions to reduce AI bias.
Show explanation
Difference: A map organises sources thematically and relationally, revealing patterns and gaps.
Show sample (synthetic)
Summary: Study A surveyed 150 teachers on AI grading tools...
Synthesis: Research on AI grading tools shows a gap between perception and performance...
Difference: Summary describes one source; synthesis integrates multiple.
Show analysis
Risk: The secondary source may have misrepresented the primary source.
What to do: Trace the claim to the original primary source and cite directly.
Show explanation
Citation laundering: Citing a source without having read it.
How AI contributes: AI may generate references that are plausible but unverified.
Show analysis
Coverage gaps: Excludes studies from non-English-speaking countries, low-income settings, and Indigenous contexts.
What to do: Search for studies from other regions and languages; acknowledge limitations.
Show explanation
How to represent conflict: Explicitly identify the disagreement, describe sources on each side, discuss possible reasons.
Example: "While Studies A and B found AI improved diagnostic accuracy, Studies C and D found no significant effect. This discrepancy may be due to differences in patient populations..."
Show TRACE analysis
- T (Trace): Search for the study in Nature.
- R (Recognise): Check for exaggeration or fabrication.
- A (Assess): Evaluate the validity — is 95% plausible?
- C (Check): Check for consistency across sources.
- E (Evaluate): Evaluate uncertainty and limitations.
Show sample response
Example (business): I would map literature on AI and supply chain management by theme, method, and region, then synthesise evidence and identify gaps.
13. Mini-Project: Verified Literature Discovery and Mapping Record
Purpose: To integrate literature discovery, verification, mapping, and synthesis skills into a structured, transparent record.
Learning outcomes: Develop search terms; distinguish discovery from verification; verify sources; construct a literature map; write a qualified synthesis.
Estimated effort: 90–150 minutes.
Available contexts: Choose one of the following research questions, or use one from Lesson 13.1 with instructor approval.
- A. Education: How do students use AI for academic writing?
- B. Health: What are the perceptions of AI-based diagnostic tools among nurses?
- C. Business: How do managers perceive the usefulness of AI for performance evaluation?
- D. Public administration: How do government communicators use AI for public messaging?
- E. Environmental studies: How are AI tools used in environmental monitoring?
Required sections
- Research question
- Core concepts
- Search terms
- Search sources
- Search dates
- Inclusion criteria
- Exclusion criteria
- AI-assisted discovery record
- Candidate source inventory
- Metadata verification
- Source authenticity
- Source relevance
- Source authority
- Source currency
- Source independence
- Primary, secondary, and tertiary sources
- Literature themes
- Methods
- Findings
- Disagreements
- Limitations
- Gaps
- Coverage limitations
- Citation-laundering concern
- Literature map
- Qualified synthesis
- Remaining uncertainty
- Human-review requirement
- AI-use disclosure
Show staged hints
Hint 1: Start with a focused question. If it is too broad, your literature search will be unmanageable.
Hint 2: Do not assume that an AI-generated reference is real. Check each one using library catalogs, Google Scholar, or DOI resolvers.
Hint 3: Organise your verified sources by theme, method, finding, and disagreement before writing your synthesis.
- I have a focused research question.
- I have developed search terms with synonyms and controlled vocabulary.
- I have verified every source I cite.
- I have removed fabricated, irrelevant, or misrepresented sources.
- I have organised my sources into a literature map.
- My synthesis is integrated and preserves uncertainty.
- I have documented my AI use and verification process.
Show partial example: Business context
Research question: "How do small-business owners perceive the usefulness of generative AI for marketing content creation?"
Core concepts: small-business owners, generative AI, marketing content, perceptions.
Search terms: "small business" OR "SME" OR "entrepreneur" AND "generative AI" OR "ChatGPT" AND "marketing" OR "content creation" AND "perception" OR "attitude."
Verified sources: 7 of 12 verified; 5 removed (3 fabricated, 2 irrelevant).
Synthesis (excerpt): Small-business owners are drawn to AI for its cost and efficiency benefits, but concerns about content quality and authenticity are widespread. Training gaps suggest that adoption may be hindered by lack of expertise. The evidence is limited to survey studies; no in-depth qualitative research has been conducted on this population.
14. Self-Check Quiz
Question 1. What is the primary distinction between literature discovery and source verification?
Show answer and explanation
Answer: B.
Question 2. Which of the following is a scholarly synthesis (not a summary or list)?
Show answer and explanation
Answer: C.
Question 3. What is selective coverage in a literature review?
Show answer and explanation
Answer: B.
Question 4. What is citation laundering?
Show answer and explanation
Answer: B.
Question 5. Which of the following is a literature map?
Show answer and explanation
Answer: C.
Question 6. What should a researcher do when an AI system provides a reference that cannot be verified?
Show answer and explanation
Answer: C.
Question 7. How does a scholarly summary differ from a literature map?
Show answer and explanation
Answer: A.
Question 8. Why is it important to identify gaps in a literature review?
Show answer and explanation
Answer: A.
15. Assessment Connection
- Assignment 1 (AI Learning and Reflection Journal): Preserve evidence of your search-term development, source verification, literature mapping, and synthesis.
- Assignment 5 (Verification and Fact-Checking): The skills you practised in Unit 11 (CLAIM) and this lesson are directly aligned.
- Assignment 4 (Applied AI Workflow Project): The literature discovery and mapping workflow can inform your workflow design project.
- Capstone (Assignment 6): A credible literature review is foundational.
16. Assignment 1 Journal Connection
- Prompt 1: How does literature discovery differ from source verification? Why is this distinction important for research integrity?
- Prompt 2: What does a literature map add beyond a simple source listing?
- Prompt 3: How does scholarly synthesis differ from an AI-generated summary? What did you have to do that the AI could not do?
- Prompt 4: What selective coverage or confirmation bias concerns arose in your literature work?
Show reflection planning checklist
- I can distinguish discovery from verification.
- I can verify a source using library tools.
- I can construct a literature map.
- I can write a qualified synthesis.
- I can identify coverage gaps and confirmation bias.
- I have documented my AI use and verification process.
17. Key Takeaways
- Discovery is not verification. AI may help you find sources, but you must verify every source independently.
- A literature list is not a literature map. A map organises sources thematically and relationally.
- A summary is not a synthesis. A synthesis integrates multiple sources into a coherent account.
- Selective coverage and confirmation bias undermine research integrity.
- Citation laundering occurs when sources are cited without being read.
- TRACE and CLAIM are essential tools for verifying literature claims and sources.
- Gaps in the literature justify new research.
- Human scholarly responsibility is non-negotiable.
18. Key Terms
- Search strategy
- A systematic plan for identifying literature, including key concepts, synonyms, controlled vocabulary, databases, inclusion and exclusion criteria, and search dates.
- Literature discovery
- The process of identifying potential sources through search engines, databases, AI, citation chaining, or other methods.
- Source verification
- The process of confirming that a source exists, the citation is accurate, and the content is represented correctly.
- Literature list
- A simple inventory of sources, such as a bibliography or reference list.
- Literature map
- A thematic or visual organisation of sources by theme, method, population, finding, disagreement, limitation, or gap.
- Scholarly summary
- A condensed account of a single source, preserving purpose, method, findings, scope, limitations, and uncertainty.
- Scholarly synthesis
- An integrated, evidence-based account of patterns, differences, debates, methods, strengths, limitations, gaps, and implications across multiple sources.
- Selective coverage
- Including only sources that support a preferred narrative, omitting contradictory or methodologically different studies.
- Confirmation bias
- The tendency to search for, interpret, and remember evidence that confirms one's pre-existing beliefs.
- Citation laundering
- Citing a source without having read it, often by copying citations from other reviews.
- Coverage audit
- A systematic review of what is included and what is missing in a literature review.
- TRACE
- A framework for identifying hallucinations, errors, omissions, invalid reasoning, and uncertainty in AI-generated content.
- CLAIM
- A verification protocol for confirming that claims are supported by evidence and accurately represented.
19. Looking Ahead
Next lesson: Lesson 13.3: AI-Assisted Analysis, Interpretation, and Research Ethics – You will learn to distinguish data preparation, analysis, interpretation, and conclusion; identify appropriate and inappropriate AI roles; verify code and calculations; and protect research data.
Unit 13 continues: Lesson 13.4 will cover RIGOUR, reproducibility, documentation, and scholarly responsibility.
After Unit 13: You will complete Module 4 and move to Module 5: AI for Professional Practice and Everyday Life.
20. Sources and Further Reading
- Primary: COMP 3270 Textbook, Chapter 14: Research and Scholarly Practice.
- Supporting: Chapter 7 (TRACE); Chapter 8 (CLAIM); Appendix C (Verification Toolkit).
- Further reading: [INSERT VERIFIED RESEARCH-METHODS REFERENCE].
- Library guides: Consult your institution's library guides on literature searching, systematic reviews, and source evaluation.
- Institutional policies: Refer to your institution's academic integrity and research ethics policies.