COMP 3270 / Assignment 4

Discipline-Specific AI Application Analysis · 15% · Module 4

Assignment 4

Evaluate one bounded generative AI application through the standards, evidence, methods, responsibilities, and human judgment of a selected discipline.

Assessment
Assignment 4
Title
Discipline-Specific AI Application Analysis
Weight
15%
Primary module
Module 4
Relevant units
Units 10–13
Core lens
Disciplinary application
Target length
2,800–3,500 words
Submission
End of Module 4

Assignment overview

Analyze one bounded generative AI application in a discipline or professional practice area. Determine where the application may be useful, what evidence and standards govern its use, what work must remain human, and what verification, accessibility, integrity, and accountability conditions are required.

Central principle: A useful AI output is not automatically a valid disciplinary product. Disciplinary quality depends on evidence, method, context, expertise, interpretation, and accountable human judgment.

Discipline-grounded

Use the standards, methods, sources, values, and authorities of the selected field.

Evidence-tested

Evaluate claims, sources, limitations, reproducibility, and validation requirements.

Human-governed

Preserve expertise, authorship, accessibility, relationships, and final authority.

What the analysis should demonstrate

  1. Define a legitimate, bounded disciplinary purpose and intended outcome.
  2. Identify applicable standards, methods, evidence, authority, and professional duties.
  3. Map the current and proposed workflow, including affected people and non-users.
  4. Compare AI-supported and non-AI alternatives.
  5. Bound the AI role and identify human-only work.
  6. Evaluate claims, sources, uncertainty, and validation requirements.
  7. Apply the most relevant disciplinary framework or frameworks.
  8. Address accessibility, inclusion, privacy, security, rights, authorship, and integrity.
  9. Assign responsibility for review, approval, correction, escalation, and stopping.
  10. Provide a qualified recommendation and residual-risk statement.

Application streams

Research and evidence

Use TRACE, CLAIM, RIGOUR, or related course concepts to examine evidence work.

Education and communication

Use LEARN, VOICE, accessibility, authorship, and audience requirements.

Professional and organizational practice

Use WORKS, LIFE, professional standards, human expertise, and accountable workflow design.

Required components

1. Use-case and discipline brief

Purpose, setting, audience, outcome, scope, standards, authority, information, and exclusions.

2. Workflow and alternatives

Current practice, proposed AI role, non-AI options, stakeholders, review gates, and human-only work.

3. Evidence and framework analysis

Claims, sources, methods, validation, uncertainty, and relevant disciplinary frameworks.

4. Responsible-practice analysis

Accessibility, inclusion, privacy, security, permission, authorship, attribution, and integrity.

5. Responsibility and safeguards

Reviewer qualifications, approval, correction, escalation, fallback, and stop authority.

6. Recommendation

Comparative judgment, implementation conditions, limitations, residual risk, and conclusion.

Recommended process

  1. Define
    Purpose, discipline, standards, evidence, people, and alternatives.
  2. Map
    Current and proposed workflows, information, AI role, and human authority.
  3. Evaluate
    Claims, sources, methods, quality, accessibility, rights, and integrity.
  4. Recommend
    Safeguards, owners, correction, stopping, residual risk, and conclusion.

Assignment 4 resources

Official Moodle due date and time: ____________________

InstructionsGrading rubricSubmission page
👤