Tutorial 3 of 5

Framework Selection and Integrated Analysis

COMP 327 Capstone Project — Assignment 6: Discipline-Specific Responsible AI Integration Strategy

Suggested completion time: 120–150 minutes

Required prior work: Tutorials 1 & 2 (scope, purpose, workflows, responsibility)

Capstone period: Weeks 12–13 (Late Week 12)

Connects to: Tutorial 4 (governance) and Tutorial 5 (synthesis)

1. Purpose of This Tutorial

This tutorial helps you select and apply relevant frameworks from the COMP 327 course to your Capstone Project. Rather than mechanically applying every framework, you will learn to match frameworks to the specific questions your project raises, and to build an integrated analysis that synthesises insights across multiple frameworks.

The course offers a rich set of frameworks (AI Paradox, ORBIT, PCAFE‑V, SCOPE, PARTNER, FAIRER, PROTECT, CREATE, TRACE, CLAIM, LEARN, RIGOUR, VOICE, WORKS, LIFE, DOMAIN, GOVERN, FUTURES, HUMAN). You will choose a primary framework that addresses your central concern, and supporting frameworks that cover specific aspects (e.g., fairness, evidence, governance). You will also identify which frameworks are not necessary and justify why.

By the end of this tutorial, you will have a framework‑selection matrix, an integrated analysis outline, an evidence map, and a residual‑uncertainty register. These products will form the analytical core of your final capstone report.

📌 Why this matters

Frameworks are not checklists; they are analytical lenses. Using them thoughtfully ensures that your strategy is grounded in established principles and that you do not overlook critical issues. A well‑selected set of frameworks makes your analysis defensible, comprehensive, and transparent.

2. Connection to the Capstone Project

Your final capstone report includes a dedicated Framework Selection and Rationale section (Section 11 of the recommended structure) and an Integrated Responsible AI Analysis (Section 12). This tutorial provides the raw material for those sections.

The frameworks you select will shape how you analyse fairness, information governance, evidence quality, human impact, and future scenarios. They also connect to the governance and safeguard design you will develop in Tutorial 4.

🔗 Links to other tutorials

Tutorial 1 defined your problem and purpose; these guide framework selection.
Tutorial 2 provided workflow and responsibility maps; frameworks like PARTNER and SCOPE directly apply to these.
Tutorial 4 will use your integrated analysis to design governance and safeguards.
Tutorial 5 will audit your analysis and ensure you have addressed all frameworks appropriately.

3. Learning Outcomes

By completing this tutorial, you will be able to:

Review the complete set of course frameworks and understand their purpose and scope.
Match frameworks to the specific questions, risks, and governance needs of your capstone project.
Distinguish between primary and supporting frameworks and justify your selections.
Explain why certain frameworks are not applicable to your project.
Apply selected frameworks to your problem, stakeholder analysis, workflow, and evidence.
Synthesise findings from multiple frameworks into a coherent, non‑repetitive integrated analysis.
Map claims to evidence sources and identify residual uncertainties.
Develop a preliminary recommendation direction based on your integrated analysis.

4. Key Concepts

This tutorial centres on framework selection, integrated analysis, and evidence mapping. Below are the key concepts you will apply.

4.1 Framework Selection Criteria

You should select frameworks based on:

Problem and purpose – What is the central challenge? Which frameworks directly address it?
Discipline and stakeholders – Which frameworks are most relevant to your professional context?
Information and AI role – What data, permissions, and AI capabilities are involved?
Workflow and responsibility – Where are the human‑AI interactions and authority points?
Level of consequence – High‑stakes contexts require more governance‑oriented frameworks (e.g., GOVERN).
Evidence required – How will you verify claims? (TRACE, CLAIM, RIGOUR).
Future uncertainties – Are you considering emerging capabilities? (FUTURES).

You are not required to apply every framework. Choose those that add value and avoid checklist‑driven use.

4.2 Primary vs. Supporting Frameworks

A primary framework is the overarching lens that shapes your entire analysis. It often addresses the core tension or opportunity in your project (e.g., AI Paradox, ORBIT, or PARTNER if allocation is central). Supporting frameworks address specific dimensions (e.g., FAIRER for fairness, GOVERN for governance).

You will typically have one primary and two to four supporting frameworks. Too many frameworks fragment the analysis; too few leave gaps.

4.3 Integrated Analysis

An integrated analysis does not present framework findings in separate, disconnected sections. Instead, it weaves them together to tell a coherent story about your proposed AI integration. You might organise by workflow steps, by stakeholder group, or by key questions, and then show how each framework sheds light on that aspect.

4.4 Evidence Map

An evidence map links every important claim in your analysis to a specific, verifiable source. This ensures that your strategy is evidence‑based and not reliant on unverified AI‑generated assertions. You will trace claims back to course materials, scholarly sources, professional standards, or other authoritative documents.

4.5 Residual‑Uncertainty Register

Even with rigorous analysis, some uncertainty will remain. A residual‑uncertainty register documents what you do not know, what assumptions you made, and where new evidence might change your conclusions. This is a hallmark of responsible AI practice.

4.6 Framework Names and Purposes (Quick Reference)

The course frameworks are grouped by theme. Use these when selecting frameworks:

Foundational: AI Paradox, ORBIT – for opportunity‑risk framing.
Prompting & Workflow: PCAFE‑V, SCOPE, PARTNER – for prompt design, context/stages, and human‑AI allocation.
Responsibility, Rights, Information: FAIRER (fairness/accessibility), PROTECT (privacy/security), CREATE (copyright/authorship).
Evidence & Learning: TRACE, CLAIM, LEARN, RIGOUR – for verifying claims, learning impact, and research integrity.
Applied Practice: VOICE (writing/communication), WORKS (organisational), LIFE (personal/consumer).
Discipline, Governance, Integration: DOMAIN, GOVERN, FUTURES, HUMAN – for translating to discipline, governance, future scenarios, and human‑centric integration.

Refer to the course materials for exact expansions; do not invent them. If you are unsure about a framework's details, use it by name and purpose only.

5. Retrieval Practice

Before you begin the guided activities, recall what you know about frameworks and integration.

1. What is the main risk of using too many frameworks in your analysis?

Think about fragmentation, repetition, and loss of focus.

2. How do you decide which framework is your primary one?

Consider what is most central: the core problem, the main tension, or the key governance need.

3. What is the purpose of an evidence map?

To link every claim to a verifiable source, ensuring accountability and credibility.

4. Why should you document residual uncertainty?

To be transparent about limitations and to guide future monitoring and adaptation.

5. Give an example of a project where PARTNER would be a primary framework.

Think of a project where human‑AI allocation and authority are the central design challenge.

6. Worked Synthetic Example

Continuing the Education stream example from Tutorials 1 and 2, the student now selects frameworks and builds an integrated analysis.

6.1 Framework‑Selection Matrix

Table 1: Framework selection for Education example
Framework	Relevance	Selected?	Role
AI Paradox	High‑level opportunity vs. risk framing	Yes	Supporting
ORBIT	Opportunity, Risk, Benefit, Impact, Trade‑offs	No (covered by AI Paradox + others)	–
PARTNER	Central: human‑AI allocation and authority	Yes	Primary
SCOPE	Context, sources, stages, evaluation gates	Yes	Supporting
FAIRER	Accessibility and inclusion for students with disabilities	Yes	Supporting
PROTECT	Privacy of lecture recordings, student data	Yes	Supporting
TRACE / CLAIM	Verification of AI‑generated summaries and claims	Yes	Supporting
GOVERN	Oversight, monitoring, incident response	Yes	Supporting
FUTURES	Adaptive safeguards for emerging capabilities	Yes	Supporting
HUMAN	Final integration of human purpose, agency, accountability	Yes	Supporting (final review)
VOICE, WORKS, LIFE	Not applicable (not writing/organisational/personal focus)	No	–
CREATE	Copyright, authorship – not central (lecture content is owned by institution)	No	–
LEARN	Learning impact – relevant but covered by FAIRER and HUMAN	No (covered)	–

Primary: PARTNER (responsibility and authority are the core design challenge).
Supporting: AI Paradox, SCOPE, FAIRER, PROTECT, TRACE/CLAIM, GOVERN, FUTURES, HUMAN.

Excluded: ORBIT (covered), VOICE, WORKS, LIFE, CREATE, LEARN (not central or covered).

6.2 Integrated Analysis Outline

The student organises the analysis around three key themes:

Human‑AI Allocation and Authority (PARTNER + SCOPE) – Map each workflow step to responsibility, review, decision, and stop authorities. Use SCOPE to define stages and evaluation gates (e.g., draft, review, approval, distribution).
Fairness, Access, and Privacy (FAIRER + PROTECT) – Assess representation of students with disabilities, ensure accessibility, and protect lecture content and student privacy through consent and data minimisation.
Evidence, Verification, and Governance (TRACE/CLAIM + GOVERN + FUTURES + HUMAN) – Plan for verifying AI outputs, establish monitoring, incident response, and adaptive safeguards for future changes. Conclude with HUMAN integration of purpose, agency, and accountability.

This structure avoids repeating framework findings; each theme draws on multiple frameworks in an integrated way.

6.3 Evidence Map (Excerpt)

Table 2: Evidence map – claim to source
Claim	Source	Verification
Students with visual impairments face delays in accessing materials.	University accessibility office survey (synthetic data); scholarly article on accessibility in higher education (cite specific).	Check survey methodology; verify article peer‑reviewed.
AI‑generated summaries require human review to ensure accuracy.	Course materials, Chapter 17; professional standards for accessibility.	Confirm source authority; check currency.
Lecture recordings are subject to institutional privacy policies.	Institutional policy document; PROTECT framework guidance.	Verify policy version; check applicable laws.

6.4 Residual‑Uncertainty Register

AI accuracy – We assume a 95% accuracy rate for summarisation; actual performance may vary.
Student adoption – We assume students will use the summaries; actual usage may be lower.
Instructor engagement – We assume instructors will review summaries; they may lack time.
Technological change – Future AI capabilities may change the viability of this approach.

6.5 Preliminary Recommendation Direction

Based on the integrated analysis, the student recommends a conditional adoption of the AI‑supported summarisation system, with a phased rollout starting with a pilot in one department, strict human review, and a clear fallback to manual transcription. The recommendation is qualified: the system should not be adopted without ongoing monitoring and a re‑evaluation after six months.

✅ Example complete

The student has thoughtfully selected frameworks, built an integrated analysis, mapped evidence, documented uncertainty, and arrived at a qualified recommendation.

7. Guided Activity

Using your project brief (Tutorial 1) and workflow/responsibility design (Tutorial 2), complete the following steps to select frameworks and plan your integrated analysis.

7.1 Review the Framework Set

Refresh your memory of all course frameworks. You can refer to the list in Section 4.6. For each framework, ask: “Does this framework address a central question, risk, or governance need in my project?”

7.2 Create a Framework‑Selection Matrix

For each framework, decide if it is Primary, Supporting, or Not Selected. Provide a brief justification. Use a table like the one in the worked example.

✏️ Your framework matrix:

List frameworks, your selection, and a short reason.

7.3 Identify Primary Framework Rationale

Write a clear rationale for your primary framework. Why is it the most appropriate lens for your analysis? What does it help you see that other frameworks might miss?

✏️ Primary framework rationale:

Explain why your primary framework is central.

7.4 Identify Supporting Framework Rationale

For each supporting framework, explain how it complements the primary framework and addresses specific aspects (e.g., fairness, evidence, governance).

✏️ Supporting framework rationales:

List each supporting framework and its specific contribution.

7.5 List Excluded Frameworks with Brief Justification

Explain why you did not select certain frameworks. This demonstrates thoughtful selection rather than random inclusion.

✏️ Excluded frameworks:

List frameworks you excluded and why.

7.6 Develop an Integrated Analysis Outline

Organise your analysis into three to five thematic sections that draw on multiple frameworks. Avoid a “framework‑by‑framework” structure. For example, themes might be: (1) Responsibility and Workflow, (2) Fairness and Access, (3) Evidence and Verification, (4) Governance and Adaptation, (5) Human Integration.

✏️ Integrated analysis outline:

Describe your thematic structure and which frameworks inform each theme.

7.7 Build an Evidence Map

For each major claim in your analysis, identify a specific source (course material, scholarly article, policy, etc.) and note how you will verify it. Use the table format from the worked example.

✏️ Evidence map:

Create a table linking claims to sources and verification methods.

7.8 Create a Residual‑Uncertainty Register

List at least three to five uncertainties, assumptions, or areas where evidence is incomplete. For each, note how it might affect your recommendation.

✏️ Residual‑uncertainty register:

Document your uncertainties and their potential impact.

7.9 Articulate Preliminary Findings and Recommendation Direction

Based on your analysis so far, what is your preliminary conclusion? Is the AI integration recommended, recommended with conditions, or not recommended? This is preliminary; you will refine it in Tutorials 4 and 5.

✏️ Preliminary findings and recommendation direction:

State your preliminary recommendation and the key findings supporting it.

8. Student Production Activity

Now produce the complete set of deliverables for this tutorial. These will form the analytical backbone of your final capstone.

📄 Required Student Products

Framework‑selection matrix – a table listing all frameworks, your selection (Primary / Supporting / Not selected), and a brief justification.
Primary‑framework rationale – a concise explanation of why your primary framework is the most appropriate.
Supporting‑framework rationales – for each supporting framework, explain its specific contribution.
List of excluded frameworks – with a brief justification for each exclusion.
Integrated analysis outline – a thematic structure with notes on which frameworks inform each theme.
Evidence map – a table linking each key claim to a source and verification method.
Residual‑uncertainty register – a list of uncertainties, assumptions, and their potential impact.
Preliminary findings and recommendation direction – a short statement of your current conclusion.

Format: You may use a word processor or spreadsheet. Ensure all products are clearly labelled and saved in your evidence‑retention folder.

⚠️ Avoid checklist‑driven use

Do not simply apply every framework mechanically. Your selection should be defensible and tailored to your project. If you find yourself struggling to justify a framework's inclusion, consider excluding it.

9. Evidence‑Retention Box

Add the following items to your evidence‑retention folder:

Framework‑selection matrix
Primary‑framework rationale
Supporting‑framework rationales
Excluded frameworks list
Integrated analysis outline
Evidence map
Residual‑uncertainty register
Preliminary findings and recommendation direction

These products will be essential for Tutorial 4 (governance) and Tutorial 5 (synthesis and audit).

10. Common Problems and Corrective Actions

Table 3: Common framework selection and integration problems
Problem	Description	Corrective Action
Applying every framework	Results in a checklist‑style analysis that lacks focus.	Select only those frameworks that are genuinely relevant. Exclude others with justification.
No primary framework	All frameworks are treated equally, leading to a flat analysis.	Identify the one framework that is most central to your project’s core challenge.
Fragmented analysis	Each framework is discussed separately without integration.	Organise by themes that cut across frameworks. Show how frameworks complement each other.
Weak evidence map	Claims are made without clear sources, or sources are not verified.	For each claim, identify a specific, authoritative source. Verify citations and dates.
Ignoring uncertainty	No residual‑uncertainty register; project appears overly confident.	Document assumptions, limitations, and areas where evidence is incomplete.
Recommendation not supported	Preliminary recommendation does not clearly follow from the analysis.	Ensure your recommendation is logically derived from the integrated analysis and evidence.

11. Responsible‑Use Boundaries

When selecting and applying frameworks, keep the following boundaries in mind:

⛔ Do Not

Do not invent or reconstruct framework expansions; use the exact author‑approved expansions from course materials.
Do not claim that a framework supports a conclusion it does not.
Do not use frameworks as a substitute for critical thinking; they are tools, not answers.
Do not present AI‑generated analysis as your own reasoning; you must integrate and verify all claims.

✅ Do

Refer to frameworks by name and purpose, and cite course materials when using them.
Justify your selections with clear reasoning linked to your project's context.
Use the evidence map to ground every claim in a verifiable source.
Be transparent about what you do not know and what you assume.

12. Three Equivalent Participation Pathways

You may complete this tutorial using any of the three pathways below.

Option A: Approved AI Tools

Use AI tools for bounded tasks such as:

Generating a candidate list of frameworks relevant to your problem (verify against course list).
Drafting an initial outline for your integrated analysis.
Suggesting potential evidence sources based on your topic.
Helping you articulate residual uncertainties.

You must critically evaluate all AI‑generated content and ensure it aligns with course frameworks and your own reasoning. You remain the author of all final products.

Option B: Supplied Capstone Records

Use supplied materials such as:

Example framework‑selection matrices for different streams.
Sample integrated analysis outlines.
Template evidence maps.
Lists of common uncertainties for your discipline.

Adapt these materials to your specific project, ensuring that your selections and analysis reflect your unique scope and purpose.

Option C: Non‑AI Project Development

Complete all activities using conventional methods:

Review course materials manually to identify relevant frameworks.
Develop your matrix and outline using word processing or spreadsheet tools.
Build your evidence map by searching library databases and professional sources.
Document uncertainties through careful reflection on your assumptions.

All pathways assess the same learning outcomes. Choose the one that best fits your workflow.

13. Accessibility Guidance

Ensure your framework selection and analysis products are accessible:

Use clear headings in your analysis outline.
Make tables (matrix, evidence map) with proper headers and a caption.
Provide a text description of any diagrams or visual representations of framework relationships.
Write in plain language, defining technical terms.
Ensure your residual‑uncertainty register is structured (e.g., bulleted list with clear categories).

14. Self‑Check Questions

1. What is the main disadvantage of using too many frameworks in your analysis?

A) It makes the analysis too long
B) It leads to fragmentation and repetition, weakening the overall argument
C) It impresses the reader with breadth
D) It is required by the assignment

Feedback: B is correct. Too many frameworks can make the analysis disjointed and repetitive. Quality over quantity is key.

2. Which framework would you use if your project's central issue is human‑AI allocation and authority?

A) FAIRER
B) PARTNER
C) PROTECT
D) TRACE

Feedback: B is correct. PARTNER is specifically designed for human‑AI allocation and authority. The others address fairness, privacy, and evidence, respectively.

3. What is the purpose of an evidence map?

A) To list all the sources you have read
B) To link each claim in your analysis to a specific, verifiable source
C) To show the relationship between frameworks
D) To provide a bibliography

Feedback: B is correct. An evidence map is a tool for accountability and verification, not just a bibliography.

4. Why should you document residual uncertainty?

A) To show that your analysis is incomplete
B) To be transparent about assumptions and limitations, and to guide future monitoring
C) To avoid making a firm recommendation
D) To reduce the word count

Feedback: B is correct. Documenting uncertainty is a sign of responsible analysis, not weakness.

5. Which of the following is not a good reason to exclude a framework?

A) The framework addresses a topic not relevant to your project
B) The framework's concerns are covered by another framework you have selected
C) You are not familiar with the framework
D) The framework would add little value to your analysis

Feedback: C is not a good reason. You should become familiar with all frameworks before deciding to exclude. If you are not familiar with one, you cannot make a justified exclusion.

6. In an integrated analysis, how should you structure your findings?

A) One section per framework
B) By thematic topics that draw on multiple frameworks
C) Alphabetically by framework name
D) In chronological order of framework development

Feedback: B is correct. Thematic structure integrates frameworks and avoids repetition.

7. What should you do if you cannot find a verifiable source for an important claim?

A) Use the claim anyway and hope no one checks
B) Mark it as an assumption and place it in the residual‑uncertainty register
C) Ask an AI to generate a source
D) Remove the claim entirely, even if it is essential

Feedback: B is correct. If you cannot verify a claim, you should explicitly identify it as an assumption and document the uncertainty. Do not fabricate sources.

8. Which framework would be most appropriate for assessing the impact of an AI system on student learning and skill development?

A) LEARN
B) WORKS
C) VOICE
D) CREATE

Feedback: A is correct. LEARN is designed for learning, teaching, and capability development. The others address organisational, communication, and copyright issues.

15. Completion Checklist

I have reviewed the purpose and connection to the Capstone Project.
I have completed the retrieval practice questions.
I have studied the worked synthetic example and understand the selection process and integrated analysis structure.
I have completed the guided activity:
- Reviewed the framework set
- Created a framework‑selection matrix
- Written primary and supporting framework rationales
- Listed excluded frameworks with justification
- Developed an integrated analysis outline
- Built an evidence map
- Created a residual‑uncertainty register
- Articulated preliminary findings and recommendation direction
I have produced all required student products.
I have saved all evidence in my evidence‑retention box.
I have reviewed the common problems and corrected any issues.
I have confirmed that I have not invented framework expansions and have used course materials.
I have chosen a participation pathway and completed the tutorial accordingly.
I have ensured my products are accessible.
I have completed the self‑check questions and reviewed the feedback.

✅ Tutorial 3 Complete

You are now ready for Tutorial 4: Governance, Scenarios, and Adaptive Safeguards. Your integrated analysis will inform the risk classification and safeguard design in the next tutorial.

16. Section Summaries

Purpose

Select and apply relevant frameworks, build an integrated analysis, and map evidence to claims.

Connection

Directly feeds into the Framework Selection and Integrated Analysis sections of your final report.

Learning Outcomes

You can now select frameworks, justify your choices, build a thematic analysis, and document uncertainty.

Key Concepts

Framework selection criteria, primary vs. supporting, integrated analysis, evidence map, residual‑uncertainty register.

Retrieval Practice

Recalled key concepts to prepare for the activities.

Worked Example

Showed a complete selection matrix, integrated outline, evidence map, and uncertainty register for an Education project.

Guided Activity

Step‑by‑step prompts to build your own framework selection and integrated analysis.

Student Production

Produced eight concrete deliverables that form the analytical core of your capstone.

Evidence‑Retention

Saved all products for later use in governance and synthesis tutorials.

Common Problems

Addressed over‑selection, fragmentation, weak evidence, and unsupported recommendations.

Responsible‑Use Boundaries

Reinforced that frameworks are tools, not answers; do not invent expansions or fabricate sources.

Participation Pathways

Offered three equivalent ways to complete the tutorial.

Accessibility

Guided you to ensure your matrix, outline, and maps are clear and structured.

Self‑Check

Tested understanding of framework selection, evidence mapping, uncertainty, and integration.

Completion Checklist

Confirmed all required work is done and no stop conditions are triggered.

📌 Looking ahead

In Tutorial 4, you will translate your integrated analysis into governance controls, risk classification, and adaptive safeguards.