Table of Contents
- 1. Lesson Overview
- 2. Learning Outcomes
- 3. Before You Begin
- 4. Retrieval Practice
- 5. Opening Scenario
- 6. Core Content
- 6.1 The CREATE Framework — Complete Integration
- 6.2 AI-Use Disclosure
- 6.3 Making Integrity Decisions
- 6.4 Integrating FAIRER, PROTECT, and CREATE
- 6.5 Correction, Escalation, and Stop Conditions
- 7. CREATE Walkthrough (Full)
- 8. Process Walkthrough
- 9. Self-Check Quiz
- 10. Practice Exercises
- 11. Homework Problems
- 12. Mini-Project
- 13. Summary & Key Takeaways
- 14. Next Steps
1. Lesson Overview
This culminating lesson for Unit 9 brings together all the concepts from the previous three lessons: source rights and permissions (9.1), human authorship and contribution (9.2), and plagiarism and academic integrity (9.3). You will now learn to apply the complete CREATE framework to analyse any AI-assisted work, prepare accurate and proportionate AI-use disclosures, and make responsible integrity decisions about whether a proposed use should proceed, be redesigned, require permission, or be stopped.
You will also integrate CREATE with the FAIRER and PROTECT frameworks to ensure that fairness, privacy, and security are considered alongside rights and integrity. This lesson prepares you directly for Assignment 3 (Responsible AI Case Analysis) and reinforces your reflective journal work for Assignment 1.
2. Learning Outcomes
By the end of this lesson, you should be able to:
- Apply the full CREATE framework to assess rights, authorship, attribution, transparency, and accountability in AI-assisted work.
- Develop an accurate and proportionate AI-use disclosure statement.
- Distinguish disclosure from permission, and explain why disclosure does not make prohibited use acceptable.
- Integrate FAIRER, PROTECT, and CREATE to identify comprehensive risks and safeguards.
- Identify correction, escalation, and stop conditions in an integrity review.
- Make and justify a final recommendation (proceed, seek permission, redesign, or stop) for a given AI-supported activity.
- Connect Unit 9 analysis to Assignments 1 and 3.
3. Before You Begin
- Assigned reading: Chapter 11 (full), and review of key sections from Chapters 6, 7, 8.
- Prerequisite knowledge: All previous Unit 9 lessons; understanding of source rights, authorship, plagiarism, and policy.
- Optional AI tool: You may use AI for practice, but do not upload copyrighted or confidential material.
- Academic-integrity reminder: This lesson emphasises responsible decision-making; it does not provide loopholes.
- Accessibility: All answers are in expandable details elements.
- Estimated preparation time: 15–20 minutes for reading and retrieval.
4. Retrieval Practice
Refresh your memory of key concepts from the previous lessons.
Question 1: From Lesson 9.1, what is the distinction between permission and attribution?
Show answer
Permission is authorisation from the rights holder to use a work; attribution is giving credit to the source. Permission does not eliminate attribution requirements, and attribution does not create permission.
Question 2: From Lesson 9.2, why can an AI system not be listed as a human author?
Show answer
AI cannot accept responsibility, verify sources, or exercise judgment—all of which are essential for human authorship. AI is a tool, not an accountable agent.
Question 3: From Lesson 9.3, what is source laundering, and why is it a concern?
Show answer
Source laundering occurs when AI-generated text obscures the original source of ideas or claims, often by fabricating or misattributing sources. It undermines integrity and can lead to misinformation.
Question 4: From Unit 7, what does the FAIRER framework’s “R” stand for, and how might it relate to integrity decisions?
Show answer
“R” stands for Representation. It reminds us to consider whether diverse perspectives and rights-holders are fairly represented, which is relevant when deciding whether to use certain sources or to attribute contributions.
5. Opening Scenario
A non-profit organisation is preparing a public report on climate change. The team uses a generative AI system to assist in drafting sections. They input several copyrighted reports (for which they have permission to quote but not to upload to a third-party service) and ask the AI to synthesise key findings. The AI produces a well-written draft, but it also fabricates a statistic and mimics the style of a prominent climate scientist. The team revises the draft, adds some personal observations, and plans to publish it with a generic acknowledgement: “AI was used in the preparation of this report.”
They have not checked the AI provider’s terms regarding uploaded content, have not verified the statistic, and have not considered whether style imitation might mislead readers. They assume that because they are a non-profit, all uses are fair.
Questions to consider:
- What are the copyright, permission, and licensing issues?
- What are the authorship and attribution issues?
- What plagiarism or integrity issues arise?
- Is the disclosure adequate? Why or why not?
- What would you recommend: proceed, seek permission, redesign, or stop?
Show guided analysis
1. Copyright/permission: The team has permission to quote, but not to upload to a third-party AI service. This likely violates the terms of use and copyright. They need permission for that specific use.
2. Authorship/attribution: The AI generated substantial content; the human contribution is minimal. The style imitation could mislead readers into thinking the scientist endorsed the report. The team should not claim sole authorship; they need to clearly separate human and AI contributions.
3. Plagiarism/integrity: The fabricated statistic is a serious integrity issue. Even with revision, the AI-generated content may contain patchwriting or source laundering. They must verify all facts.
4. Disclosure: “AI was used” is vague and insufficient. They need to specify what the AI did, which parts were AI-generated, and what human revisions were made. They also need to address the style imitation.
5. Recommendation: This should not proceed as is. They should: obtain permission for uploading, verify all statistics, rewrite sections to avoid style imitation, and produce a detailed disclosure. If they cannot obtain permission, they should not upload the copyrighted reports. They might redesign by summarising manually.
6. Core Content
6.1 The CREATE Framework — Complete Integration
CREATE is a structured approach to analysing rights, authorship, attribution, transparency, and accountability in AI-assisted work. The framework, as defined in Chapter 11, consists of six elements (the exact author-approved expansion is provided below). We have already explored the first four elements across Lessons 9.1–9.3. Now we bring them together and add the final two.
Operationally, CREATE addresses:
- C – Source and rights identification
- R – Permission and licence conditions
- E – Human creative and intellectual contribution
- A – Authorship and attribution
- T – Transparency and AI-use disclosure
- E – Evaluation, verification, responsibility, and correction
When applying CREATE, you systematically examine each element to ensure that all aspects of responsible use are considered. It is not a checklist for legal compliance but a guide for ethical decision-making.
6.2 AI-Use Disclosure
AI-use disclosure is a formal, truthful statement that identifies how AI was used in the creation of a work, including the system, purpose, tasks, outputs, human revisions, verification, and final responsibility.
An effective disclosure should include:
- AI system(s) used (name, version if known).
- Dates or period of use.
- Purpose of using AI.
- Tasks supported (e.g., brainstorming, drafting, editing, data analysis, image generation).
- Which outputs were used and how they were incorporated.
- Human revision and editing performed.
- Verification steps taken (e.g., fact-checking, source verification).
- Limitations of the AI-generated content (e.g., known errors, uncertainties).
- Final human responsibility statement.
What to avoid:
- Vague statements: “AI was used.” “Edited with technology.”
- Overly broad claims: “AI assisted with everything.”
- Including confidential prompt content in public disclosures.
Example: Detailed vs. Vague Disclosure
Vague: “I used AI to help write this paper.”
Detailed: “I used ChatGPT (OpenAI, GPT-4) to generate an initial outline and a first draft of the literature review. I then revised the content extensively, added my own analysis, verified all citations, and restructured the argument. I take full responsibility for the final work.”
6.3 Making Integrity Decisions
After applying CREATE and considering all aspects, you must make a decision. Possible recommendations include:
- Proceed – all requirements met, low risk.
- Proceed with attribution and disclosure – necessary to meet integrity standards.
- Proceed after verifying the licence – uncertainty about rights needs resolution.
- Obtain permission – rights holder must authorise use.
- Limit or replace source material – reduce risk by using less protected content.
- Redesign the AI role – change how AI is used to avoid substitution or integrity issues.
- Increase human authorship and revision – ensure the human contribution is significant.
- Use a non-AI alternative – avoid AI risks entirely.
- Require academic-integrity, copyright, or specialist review – seek expert guidance.
- Pause – temporarily halt while gathering more information.
- Do not proceed – the risks are too high.
- Withdraw or correct the work – if already submitted, correct errors and retract if necessary.
Your decision should be justified based on the evidence from your analysis.
6.4 Integrating FAIRER, PROTECT, and CREATE
Responsible AI use requires a holistic view. CREATE focuses on rights, authorship, and integrity. It should be used alongside other frameworks:
- FAIRER (Fairness, Accountability, Inclusivity, Representation, Equity, Rights) – ensures that the use respects diversity, representation, and equity.
- PROTECT (Privacy, Rights, Ownership, Terms, Ethics, Confidentiality, Trust) – addresses privacy, security, and data protection.
When analysing a case, ask:
- Are there fairness or inclusivity concerns? (FAIRER)
- Are there privacy or security risks? (PROTECT)
- Are rights, attribution, and integrity addressed? (CREATE)
- What shared escalation or remedy needs arise?
For example, uploading a dataset might involve copyright (CREATE), privacy of individuals (PROTECT), and potential bias in the data (FAIRER). All must be considered.
6.5 Correction, Escalation, and Stop Conditions
Correction conditions: If errors are found (e.g., fabricated citations, misattributed ideas), you must correct them promptly. This may involve revising the work, issuing a correction, or retracting a publication.
Escalation conditions: When you cannot resolve an issue on your own, escalate to a supervisor, instructor, legal counsel, or other appropriate authority. Escalation is required when:
- Source ownership is unclear.
- Licence terms are ambiguous.
- Permission is absent for substantial content.
- A living creator’s distinctive style is requested.
- Authorship is disputed.
- AI use may be prohibited.
- Citations may be fabricated.
- Confidential sources were uploaded.
- Synthetic media may mislead.
- An identifiable person may be impersonated.
Stop conditions: You must stop immediately if:
- AI use is explicitly prohibited.
- Permission is absent and cannot be obtained.
- A quotation is unverified.
- A source is fabricated.
- An authorship claim is misleading.
- Material AI contribution is concealed.
- An ownership dispute is unresolved.
- A request to imitate or impersonate is made.
- You cannot identify the source origin.
- You cannot verify affected rights.
- You cannot correct misleading provenance.
- Applicable policy prohibits continuation.
Stopping is a responsible action when integrity is at risk.
7. CREATE Walkthrough (Full Framework)
Below is a complete walkthrough of the CREATE elements, with definitions, guiding questions, examples, common mistakes, safeguards, escalation/stop conditions, and residual limitations.
Element C: Source and Rights Identification
Identify the work, its creator, the rights holder, and the applicable rights.
- Guiding questions: What is the work? Who created it? Who holds the rights?
- Example: A photograph on a public website. Rights holder: photographer.
- Common mistake: Assuming public availability means no rights.
- Safeguard: Search for copyright notices and licence information.
- Escalation: If rights holder cannot be identified, do not proceed.
- Residual limitation: Rights may be complex; multiple layers of rights may exist.
Element R: Permission and Licence Conditions
Determine whether permission or a licence applies, and identify the conditions.
- Guiding questions: Is there a licence? What are its conditions? Does your use align?
- Example: CC BY-NC 4.0 – non-commercial, attribution required.
- Common mistake: Assuming attribution is enough, ignoring other conditions.
- Safeguard: Read the full licence text.
- Stop condition: If the licence does not permit your use, stop.
- Residual limitation: Licence interpretation can be ambiguous; seek advice.
Element E: Human Creative and Intellectual Contribution
Identify the human contributions: conception, selection, analysis, argument, expression, revision, verification, approval.
- Guiding questions: What did the human do intellectually? What did AI do?
- Example: Human designed study, wrote manuscript; AI provided grammar suggestions.
- Common mistake: Overstating human contribution when AI did most work.
- Safeguard: Document contributions clearly.
- Escalation: If human contribution is minimal, redesign.
- Residual limitation: Thresholds vary by discipline.
Element A: Authorship and Attribution
Determine who should be listed as author(s), and provide appropriate attribution and acknowledgement.
- Guiding questions: Who made significant contributions? Who is accountable?
- Example: Two researchers are co-authors; a research assistant is acknowledged.
- Common mistake: Listing AI as a co-author.
- Safeguard: Follow institutional and publication authorship criteria.
- Stop condition: If authorship is disputed, pause and resolve.
- Residual limitation: Norms vary by field.
Element T: Transparency and AI-Use Disclosure
Provide accurate, proportionate disclosure of how AI was used.
- Guiding questions: What AI system? What tasks? What outputs used? What revisions?
- Example: Detailed statement as shown earlier.
- Common mistake: Vague or incomplete disclosure.
- Safeguard: Be specific and truthful.
- Stop condition: If you cannot disclose honestly without revealing prohibited conduct, stop.
- Residual limitation: Some contexts may restrict disclosure of proprietary prompts.
Element E: Evaluation, Verification, Responsibility, and Correction
Verify all content, sources, and claims; take responsibility; correct errors.
- Guiding questions: Have you verified facts? Are you prepared to defend the work?
- Example: Fact-checked all statistics, checked citations against original sources.
- Common mistake: Assuming AI-generated content is accurate.
- Safeguard: Create a verification record.
- Escalation: If you cannot verify, seek help or redesign.
- Residual limitation: Verification can be time-consuming; prioritise critical claims.
8. Process Walkthrough: Complete Rights and Integrity Review
Follow this process to systematically review a proposed AI-supported activity and make a final decision.
| Stage | Action | Guiding Question | Example |
|---|---|---|---|
| 1. Source & Rights | Identify all sources, creators, and rights holders. | What sources are used? Who owns them? | Three journal articles, two images. |
| 2. Permission & Licence | Check licences, permissions, and terms. | Are the uses permitted? | Articles under open access, images CC BY. |
| 3. Human Contribution | Document human intellectual contributions. | What did the human do? | Researcher designed study, wrote introduction. |
| 4. AI Contribution | Document AI tasks and outputs. | What did AI generate? | AI drafted the methods section. |
| 5. Authorship & Attribution | Determine authors and acknowledgements. | Who is accountable? | Researcher is sole author, AI acknowledged. |
| 6. Plagiarism & Paraphrasing | Check for copying, patchwriting, source laundering. | Is the expression original? | Review AI output for similarity to sources. |
| 7. Style & Synthetic Content | Evaluate style imitation and synthetic media. | Does it mislead? | No style imitation; synthetic images labelled. |
| 8. Assistance vs. Substitution | Assess whether AI use is appropriate. | Is this substitution? | AI only for editing, not core content. |
| 9. Disclosure | Prepare a specific disclosure statement. | What disclosure is needed? | Detailed disclosure as per example. |
| 10. Verification & Responsibility | Verify all claims and accept responsibility. | Is everything verified? Are you accountable? | All data verified; researcher accepts responsibility. |
| 11. Corrections & Escalation | Plan for corrections and escalation if needed. | What if errors are found? | Process for updating and correcting identified. |
| 12. Final Decision | Make a recommendation. | Proceed, redesign, seek permission, or stop? | Proceed with disclosure. |
Complete walkthrough example
Scenario: A student wants to use an AI to generate a summary of a journal article for a class discussion. The article is accessible through the library. The student will not share the summary widely.
- Source: Article, copyright held by publisher.
- Permission: Library access allows reading, but uploading to AI may violate terms.
- Human: Student reads article, takes notes, decides what to summarise.
- AI: Generates a summary based on the uploaded text (if permitted).
- Authorship: Student is author of the summary, but must cite article.
- Plagiarism: Summary must be in student’s own words; check for patchwriting.
- Style: No style imitation.
- Assistance: Using AI for summary may be substitution if the assignment requires the student to demonstrate comprehension.
- Disclosure: Disclose AI use.
- Verification: Check summary for accuracy against article.
- Correction: If errors, revise.
- Decision: Given that uploading may violate terms and the assignment likely expects personal summary, the student should not upload; instead, write the summary themselves, possibly using AI only for grammar. Recommendation: Redesign – write summary manually.
9. Self-Check Quiz
Question 1: What is the primary purpose of an AI-use disclosure?
Show answer and explanation
Answer: b.
The primary purpose of an AI-use disclosure is to provide transparency about how AI was used, not to justify or grant permission.
Question 2: Which of the following is not a required element of an effective AI-use disclosure?
Show answer and explanation
Answer: c.
Disclosures should not include confidential prompts; they should focus on the nature of AI use, not verbatim prompts.
Question 3: True or False: Disclosure of AI use can make prohibited AI use acceptable.
Show answer and explanation
Answer: False.
Disclosure does not cure prohibited use; it is a transparency measure, not a permission-granting mechanism.
Question 4: Which framework is most directly concerned with privacy and data protection?
Show answer and explanation
Answer: c.
PROTECT addresses privacy, security, and data protection.
Question 5: In the CREATE framework, which element addresses verification and correction of errors?
Show answer and explanation
Answer: d.
E (Evaluation, verification, responsibility, and correction) addresses verification and correction.
Question 6: Which of the following is a stop condition?
Show answer and explanation
Answer: b.
A citation that appears to be fabricated is a clear stop condition because it undermines integrity and cannot be used without verification.
Question 7: When making an integrity decision, what is the appropriate response if you find that a source is copyrighted and you do not have permission to upload it to an AI system?
Show answer and explanation
Answer: d.
You should seek permission or use an alternative, and redesign the AI role to avoid the problematic upload.
Question 8: Which of the following is an example of a good AI-use disclosure?
Show answer and explanation
Answer: b.
Option (b) is specific and accurate. The others are vague or misleading.
10. Practice Exercises
Apply CREATE, disclosure, and decision-making.
- Difficulty: Intermediate
- Estimated time: ~15 min
- Learning outcome: Apply the full CREATE framework.
Scenario: A student uses an AI to generate a visual summary of a complex concept for a class presentation. The student provides the AI with a text description and asks it to create an infographic. The student reviews and adjusts the infographic, adds their own commentary, and presents it.
Questions:
- Apply the CREATE framework to this scenario: what are the key considerations for each element?
- What disclosure would you recommend?
Show sample answer
C: The source is the student’s text description; if it incorporates others’ ideas, those need attribution.
R: The AI tool’s terms may affect usage; check for restrictions on generated content.
E: Human provided description and adjustments; AI generated the visual elements.
A: Student is author of the commentary and concept; AI is a tool.
T: Disclose that AI was used to generate the infographic, specify the tool.
E: Verify that the infographic accurately represents the concept and correct any errors.
Disclosure: “I used [AI tool] to create the infographic based on my description. I reviewed and adjusted the output. The commentary and final presentation are my own.”
- Difficulty: Intermediate
- Estimated time: ~12 min
- Learning outcome: Document AI and human contributions.
Scenario: You are writing a research paper. You used AI to:
- Brainstorm potential research questions.
- Generate a first draft of the literature review.
- Check grammar and style.
You then heavily revised the literature review, added your own analysis, and verified all sources.
Question: Create an AI contribution record that documents these uses.
Show sample answer
AI Contribution Record
- System: ChatGPT (OpenAI, GPT-4).
- Dates: March 10–15, 2026.
- Tasks: Brainstorming research questions (March 10), generating a first draft of the literature review (March 12), grammar and style suggestions (March 14).
- Outputs used: The brainstormed list and the draft literature review were used as starting points. Grammar suggestions were applied.
- Human revisions: I substantially revised the literature review, reorganising it, adding critical analysis, and writing new sections. I also added my own research questions and conclusions.
- Verification: I checked all citations against original sources and verified all factual claims.
- Final responsibility: I take full responsibility for the final paper.
- Difficulty: Introductory
- Estimated time: ~10 min
- Learning outcome: Develop an effective AI-use disclosure.
Scenario: You used an AI to create a data visualisation for a report. You provided the data and specified the chart type. The AI generated the chart, and you adjusted the labels and colours.
Question: Write a disclosure statement for the report.
Show sample answer
Disclosure: “I used [AI tool] to generate the initial chart for this report. I provided the underlying data and specified the chart type. I subsequently adjusted the labels, colours, and layout to improve clarity. I take responsibility for the accuracy and interpretation of the data presented.”
- Difficulty: Intermediate
- Estimated time: ~10 min
- Learning outcome: Evaluate and improve disclosures.
Inadequate disclosure: “AI was used in the preparation of this work.”
Questions:
- Why is this disclosure inadequate?
- Rewrite it to be more effective.
Show sample answer
Why inadequate: It is vague; it does not identify the tool, the nature of AI assistance, or the human contributions. It could apply to anything from grammar checking to full generation, leaving the reader uncertain.
Rewritten: “I used Grammarly to check grammar and spelling in the final draft. I also used ChatGPT (OpenAI, GPT-4) to generate a preliminary outline and to suggest alternative phrasings for a few sections. All AI-generated suggestions were reviewed and significantly revised by me. I take full responsibility for the final content and its accuracy.”
- Difficulty: Introductory
- Estimated time: ~8 min
- Learning outcome: Distinguish disclosure from permission.
Scenario: A student discloses that they used AI to generate a full essay, and they also cite the AI as a source. However, the assignment explicitly forbids AI use for this task.
Question: Does the disclosure make the use acceptable? Why or why not?
Show sample answer
No. Disclosure does not override the assignment’s prohibition. The student violated the rule; disclosure merely makes the violation transparent, but it does not make it permissible. The student should not have used AI for that task.
- Difficulty: Advanced
- Estimated time: ~15 min
- Learning outcome: Integrate the frameworks.
Scenario: A company plans to use an AI to analyse customer feedback data to improve product design. The data includes demographic information (age, gender, location) and free-text comments. The AI will be used to identify common themes and sentiment. The company wants to publish a summary of findings.
Questions:
- What are the FAIRER concerns (fairness, representation)?
- What are the PROTECT concerns (privacy, confidentiality)?
- What are the CREATE concerns (rights, attribution, integrity)?
- What is your overall recommendation?
Show sample answer
FAIRER: The data may be biased if certain demographics are underrepresented. Representation of diverse customer voices is important. The analysis should be fair and not perpetuate stereotypes.
PROTECT: Customer feedback may contain personal information. The company must ensure anonymisation, comply with privacy regulations (e.g., GDPR), and obtain consent for analysis.
CREATE: The feedback is owned by customers or the company? The AI-generated summary must attribute ideas to customers appropriately. The company must verify the accuracy of the AI analysis (no hallucinated themes). Disclosure of AI use in the summary is needed.
Recommendation: Proceed with safeguards: anonymise data, obtain necessary consents, verify AI output, and disclose AI use. Ensure the summary is fair and representative.
- Difficulty: Advanced
- Estimated time: ~15 min
- Learning outcome: Make a final integrity recommendation.
Scenario: A graduate student is writing a thesis. They have used AI to generate a first draft of their literature review. They have also used AI to suggest citations. Some of the suggested citations appear to be fabricated. The student has not disclosed AI use. The thesis is due in a week.
Questions:
- Identify the key issues.
- What should the student do?
- Make a final recommendation.
Show sample answer
Issues: Fabricated citations (stop condition), undisclosed AI use (disclosure lacking), potential patchwriting or lack of human contribution.
Should do: Immediately stop using the AI-generated content with fabricated citations. Verify every citation against real sources. Rewrite the literature review independently. Disclose AI use for brainstorming/editing only, not for generating content. If the student cannot verify or rewrite in time, they should request an extension and seek advice from their supervisor.
Recommendation: Do not submit the thesis with fabricated citations. Withdraw the current version, correct it, disclose appropriately, and seek guidance. If necessary, delay submission.
11. Homework Problems
Match each CREATE element to its primary focus: (a) Source identification, (b) Permission and licence, (c) Human contribution, (d) Authorship and attribution, (e) Transparency, (f) Verification and correction.
Show solution design
Solution: C – Source identification; R – Permission; E (first) – Human contribution; A – Authorship; T – Transparency; E (second) – Verification and correction.
A journalist uses an AI to summarise a public report and then writes an article based on the summary. The journalist attributes the report but does not mention AI. Apply CREATE to this case and identify gaps.
Show solution design
Solution: C – report is public; R – report may be openly licensed; E – human provided direction and writing; A – journalist is author; T – missing disclosure of AI; E – verification of summary against report needed. Recommendation: disclose AI use and verify accuracy.
You used AI to translate a document from Spanish to English, then edited the translation for clarity. Write a disclosure.
Show solution design
Solution: “I used [AI translation tool] to generate an initial translation from Spanish to English. I then edited the translation for accuracy, clarity, and style. I take responsibility for the final English version.”
Disclosure: “AI was used for editing.” Rewrite it to be sufficient, given that the AI also generated a first draft.
Show solution design
Solution: “I used ChatGPT to generate a first draft of the introduction. I then heavily revised the draft, rewriting many sections and adding my own analysis. I also used Grammarly for final grammar checks. I take full responsibility for the final content.”
A company discloses that they used AI to generate product descriptions, but the images used in the descriptions were copyrighted and used without permission. Does disclosure address the copyright issue?
Show solution design
Solution: No. Disclosure of AI use does not address the unauthorised use of copyrighted images. The company must obtain permission or use properly licensed images.
You are writing a policy analysis. You used AI to gather initial data points and to create a summary table. You verified all data and wrote the analysis yourself. Create a provenance record for this work.
Show solution design
Solution: List all sources used, note that AI was used to extract data points and generate a table, provide the prompts used, document the verification of each data point against original sources, and state that the analysis and writing are your own.
A university uses an AI to screen student applications for a scholarship. The AI analyses essays and recommends candidates. The university does not disclose the AI use to applicants. What are the issues from each framework?
Show solution design
FAIRER: The AI may have biases that disadvantage certain groups; representation and fairness are concerns.
PROTECT: Student essays contain personal information; privacy must be protected.
CREATE: The AI’s role is not disclosed (T); the university must verify the AI’s recommendations (E).
Recommendation: The university should disclose AI use, ensure fairness, protect privacy, and verify recommendations.
After submitting a report, you discover that an AI-generated statistic was incorrect. What should you do?
Show solution design
Solution: Immediately correct the statistic in the report, issue a correction to all recipients, and update any online versions. Apologise and explain the error. This demonstrates responsibility.
You are using an AI to analyse a dataset and you are unsure whether the dataset is copyrighted or whether you have permission to use it. What should you do?
Show solution design
Solution: Do not proceed with the analysis. Escalate the issue to your supervisor, the data provider, or legal counsel to determine the rights status. Obtain clarification before continuing.
You are using an AI to generate a business report. The AI produces text that appears to be copied from a competitor’s website, and you cannot verify the originality. What should you do?
Show solution design
Solution: Stop using the AI-generated text immediately. Do not include it in the report. Instead, write original content or obtain permission to use the competitor’s material. If necessary, redesign the AI role.
A student wants to use an AI to generate a summary of a lecture for personal study. The lecture is not copyrighted, and the student will not share the summary. The student wants to use the AI to save time. What is your recommendation?
Show solution design
Solution: Proceed. This is low-risk: the lecture is not copyrighted, the use is personal, and no publication is involved. The student may still want to verify the summary for accuracy and consider whether they are learning effectively by using the AI.
You are taking over a project from a colleague who used AI extensively but did not document their use. You need to finalise the project. What steps do you take?
Show solution design
Solution: Review the work to identify possible AI-generated sections. Verify sources and claims. If you cannot determine the origin, consider rewriting those sections or disclosing the uncertainty. Discuss with your supervisor or team to decide on authorship and disclosure.
How would you apply CREATE to your chosen case for Assignment 3? List the key questions you would ask for each element.
Show solution design
Solution: For C: What are the sources and who owns them? For R: What licences or permissions apply? For E: What are the human contributions? For A: Who is the author? For T: What AI use must be disclosed? For E (second): How will you verify and take responsibility? This framework will guide your case analysis.
12. Mini-Project: Complete CREATE Integrity Review
- Purpose: Conduct a comprehensive integrity review of a synthetic AI-assisted work, applying all Unit 9 concepts and the full CREATE framework.
- Learning outcomes: All Unit 9 outcomes, with emphasis on rights, authorship, integrity, disclosure, and decision-making.
- Estimated effort: 90–150 minutes.
Scenario
You are a consultant hired by a health non-profit to produce a report on “Nutritional Challenges in Low-Income Communities.” The report will be published on their website and shared with policymakers.
You used an AI system in the following ways:
- Data collection: You asked the AI to compile statistics from various public health databases. The AI provided a table of numbers with sources.
- Drafting: You asked the AI to write a first draft of the report’s background section, based on your outline.
- Image generation: You used an AI image generator to create a cover image for the report. You prompted it with “healthy food in a low-income community.”
- Editing: You used an AI grammar checker on the final draft.
You then:
- Revised the background section substantially, adding your own analysis and local examples.
- Verified the statistics against the original sources provided by the AI. You found that two numbers were incorrect.
- Kept the cover image as generated.
- Did not disclose any AI use.
- Checked the licences of the databases mentioned by the AI and found they were all open access.
You are now ready to finalise and publish the report.
Required Sections
Complete the following 47 sections. For brevity in this lesson, we provide a condensed version; however, in your own work, you should address each point thoroughly.
- Use-case title: Nutritional Challenges Report.
- Human problem: Describe the problem the report addresses.
- Legitimate purpose: Why is the report needed?
- Intended audience: Who will read it?
- AI role: What did the AI do?
- Source inventory: List all sources (databases, images).
- Creator and rights-holder questions: Who owns the sources?
- Access conditions: How were sources accessed?
- Permission evidence: What permissions exist?
- Licence conditions: What licences apply?
- Upload authorisation: Was uploading to AI permitted?
- Proposed transformation: How did you transform the sources?
- Attribution requirements: What attribution is needed?
- Citation requirements: What citations are needed?
- Human conception: What ideas did you bring?
- Human analysis: What analysis did you do?
- Human creative contribution: What creative choices did you make?
- AI contribution: What did AI contribute?
- Human revision: How did you revise AI output?
- Verification: What verification steps did you take?
- Authorship designation: Who is the author?
- Contributor acknowledgement: Who should be acknowledged?
- Originality analysis: Is the work original?
- Verbatim-copying review: Any verbatim copying?
- Paraphrasing review: Is paraphrasing responsible?
- Patchwriting review: Any patchwriting?
- Structural-imitation review: Does structure imitate another?
- Source-laundering review: Any source laundering?
- Style-imitation analysis: Any style imitation?
- Synthetic quotation or media analysis: How are synthetic media labelled?
- Assistance-versus-substitution analysis: Is AI use assistance or substitution?
- Applicable academic or professional policy: What policies apply?
- Provenance record: What records do you have?
- AI-use disclosure: Draft your disclosure.
- Disclosure sufficiency: Is it sufficient?
- Fairness and accessibility interaction: Any fairness issues?
- Privacy and security interaction: Any privacy issues?
- Human responsibility statement: Who accepts responsibility?
- Correction process: How would you correct errors?
- Escalation: What would trigger escalation?
- Stop conditions: What would stop you?
- Complete CREATE analysis: Summarise CREATE findings.
- Residual uncertainty: What remains unclear?
- Final recommendation: Proceed, redesign, etc.
- Assignment 3 connection: How does this prepare you for Assignment 3?
- Assignment 1 reflection connection: How does this inform your journal?
- Preparation for Unit 10: What verification methods will you need?
Show staged hints
Hint 1: For source rights, the databases are open access, but the AI-generated image may have its own licensing from the AI provider.
Hint 2: The two incorrect statistics are a verification issue; you must correct them.
Hint 3: Lack of disclosure is a major gap; you need to draft a detailed statement.
Hint 4: The cover image may raise synthetic media labelling concerns.
Hint 5: The AI helped draft background, but you revised substantially; assess whether you can claim authorship.
Show solution architecture
- Use-case: Report on nutritional challenges.
- Purpose: Inform policy and public.
- AI role: Data compilation, drafting, image generation, editing.
- Sources: Public health databases (open access), AI-generated image.
- Permissions: Databases are open; image may have terms from AI provider.
- Human contribution: Outlined, revised, added analysis, verified data.
- AI contribution: Compiled data (with errors), draft, image.
- Verification: Corrected two errors; need to verify all data.
- Authorship: You are the author, but AI acknowledged.
- Disclosure: Draft detailed disclosure.
- Ethics: Ensure image is labelled as AI-generated.
- Recommendation: Proceed with corrections and disclosure.
Show illustrative partial example (selected sections)
Section 11: Upload authorisation: The databases are open access and allow use, but the AI provider’s terms may restrict commercial use of generated images. Check terms.
Section 18: AI contribution: AI compiled data (with errors), drafted background, generated cover image.
Section 20: Verification: I checked all statistics against original sources and corrected two errors. I also verified that the databases were correctly cited.
Section 34: AI-use disclosure: “I used [AI tool] to compile preliminary statistics, generate a first draft of the background section, and create the cover image. I revised the draft extensively, verified all data, and corrected errors. The cover image is AI-generated and labelled accordingly. I take full responsibility for the report’s accuracy and conclusions.”
Section 41: Stop conditions: If I find additional fabricated data or cannot verify a source, I will stop and rewrite.
Section 44: Final recommendation: Proceed with the report, incorporating the corrections and the disclosure statement.
Self-Review Checklist
- Did I address all 47 sections?
- Did I apply CREATE to each aspect?
- Did I correct all errors and verify all sources?
- Is my disclosure specific and accurate?
- Did I make a justifiable final recommendation?
13. Summary & Key Takeaways
- CREATE is a comprehensive framework for analysing rights, authorship, attribution, transparency, and accountability.
- AI-use disclosure must be specific, truthful, and proportionate; it does not cure prohibited use.
- Integrity decisions range from proceed to stop, based on careful analysis of all risks.
- Integrate FAIRER (fairness, representation) and PROTECT (privacy, security) with CREATE for holistic review.
- Know your correction, escalation, and stop conditions and act on them responsibly.
- Always connect your analysis to applicable policies and seek advice when uncertain.
Common Misconceptions
- ❌ “CREATE is a legal checklist.” → ✅ It is a guide for ethical decision-making, not a substitute for legal advice.
- ❌ “Disclosure is enough.” → ✅ Disclosure is important, but the use must also be permitted and ethical.
- ❌ “Once I disclose, I’m done.” → ✅ You must also verify, correct, and take responsibility.
- ❌ “If the AI did it, it’s not my fault.” → ✅ You are responsible for the final work.
14. Next Steps
Congratulations! You have completed Unit 9. You now have a solid foundation in copyright, authorship, attribution, academic integrity, and the CREATE framework for responsible AI use.
Assignment 3: Use the CREATE framework and the comprehensive review process from this lesson to analyse your chosen case study. Your analysis should cover all relevant aspects of rights, authorship, attribution, integrity, and disclosure.
Assignment 1: Reflect on how your understanding of responsible AI use has evolved throughout Unit 9. Consider the journal prompts provided in the Unit 9 master prompt.
Next unit: Unit 10 will focus on verification, evidence, and epistemic responsibility, including detailed methods for detecting fabricated citations and evaluating claims.