7 The AI Critique Toolkit - Becoming a Smart Business Professional

The most dangerous AI output is not the one that is obviously wrong. It is the one that sounds exactly right, and nobody checks.

7.1 Why Business Students Need Critique Skills

Imagine you just hired a brilliant but inexperienced analyst or consultant. They work fast, have lots of ideas, but sometimes:

Overcomplicate simple solutions
Miss important legal, technical, or operational considerations
Write recommendations that sound good but have practical flaws
Make assumptions about your workplace context, capacity, or constraints

This is exactly how AI behaves. Your job as a business professional is to review, question, and improve AI-generated advice before it impacts real people and real organisations.

The difference between good business professionals and great ones isn’t whether they use AI tools—it’s how critically they evaluate AI outputs. In professional work across all disciplines, bad advice can lead to legal challenges, damaged relationships, failed initiatives, or strategic missteps.

7.2 The Business Impact of Uncritical AI Acceptance

Scenario Examples Across Disciplines:

Example: Management

Your AI generates a change management plan. You follow it without stakeholder testing. Outcomes:

Key influencers weren’t engaged early enough
Communication messaging doesn’t resonate with your culture
Implementation timeline is unrealistic given other priorities
Change initiative stalls after initial enthusiasm

Lesson: Always critique AI-generated recommendations thoroughly before implementation, regardless of discipline.

7.3 Your 5-Step Business Critique Framework

This framework applies across all business disciplines:

Comprehension Check - “Do I understand this completely?”
Simplicity Check - “Is this practical for my workplace/situation?”
Legal, Technical & Risk Check - “What are the legal, technical, ethical, and operational implications?”
Context Check - “What assumptions is AI making about my organisation?”
Stakeholder Check - “How will different groups react to this?”

7.4 Step 1: Comprehension Check - “Do I understand this?”

Red Flags: - HR jargon that sounds impressive but is unclear - Policy language that could be interpreted multiple ways - No clear explanation of why this approach is recommended

Your Response: - “Can you explain this in plain English that a line manager could understand?” - “What specific problem does this solve and how does it solve it?” - “Break this down into step-by-step actions that need to be taken”

Example:

# ❌ AI gives you this:
"We recommend implementing a holistic performance ecosystem leveraging"
"synergistic feedback loops and agile recalibration mechanisms."

# ✅ Ask for this instead:
"We recommend creating a performance management system where employees get feedback from multiple sources throughout the year, not just once annually. This includes regular check-ins with managers, peer feedback on projects, and quarterly progress reviews against clear goals."

7.5 Step 2: Simplicity Check - “Is this practical for my workplace?”

Red Flags: - Solution requires resources you don’t have (extra HR staff, expensive software) - Process is more complex than your current problem - Assumes perfect implementation with no room for human error

Your Response: - “Give me a version that works with a 2-person HR team and limited budget” - “What’s the minimum viable version of this solution?” - “Show me how to implement this step by step over 6 months, not all at once”

HR Example: AI suggests a sophisticated 360-degree feedback system with custom software, external facilitators, and detailed analytics. You ask for a simpler approach and get a practical solution using existing tools like Google Forms and manager training sessions.

7.6 Step 3: Legal, Technical & Risk Check - “What are the implications?”

Red Flags (vary by discipline): - No mention of compliance, legal, technical, or regulatory considerations - Recommendations that could harm certain groups (employees, customers, stakeholders) - Risk implications that aren’t addressed - No consideration of organisational constraints or industry requirements

Your Response (discipline-specific examples):

Example: Supply Chain

“What are the operational and financial risks?”
“How does this affect compliance and sustainability?”
“What supply chain resilience issues exist?”

Critical Questions to Always Ask (across all disciplines): - What could go wrong and what’s the exposure? - Are we meeting our compliance and governance obligations? - What organisational constraints or capabilities might we lack? - Who might be negatively affected and how?

7.7 Step 4: Context Check - “What assumptions is AI making about my organisation?”

Red Flags: - AI assumes you have a sophisticated HR information system - Assumes your managers are all skilled in difficult conversations - Assumes your workplace culture is collaborative and high-trust - Assumes employees are digitally literate and open to change

Your Response: - “What assumptions are you making about our current systems and processes?” - “How would this work in a workplace with high employee turnover?” - “What if our managers resist this change?” - “How does this account for our hybrid/remote/in-person work arrangement?”

Real Example: AI recommends an employee engagement app with daily mood tracking and gamified recognition. You realise this assumes your employees are comfortable sharing personal data and that you have the technical infrastructure to support it. You ask for alternatives that work with your existing communication channels.

7.8 Step 5: Stakeholder Check - “How will different groups react to this?”

Red Flags: - Solution only considers management perspective - No thought to how employees will perceive or experience the change - Doesn’t address how different departments might be affected differently - No consideration of change management requirements

Your Response: - “Walk me through how this would feel from an employee’s perspective” - “How might different departments (sales, operations, finance) experience this differently?” - “What resistance should we anticipate and how can we address it?” - “What communication and training would be needed for successful implementation?”

Stakeholder Mapping (adjust by discipline): Always consider:

Senior Leadership: Will they see the business value and strategic alignment?
Operational Staff/Managers: Do they have skills, capacity, and buy-in to implement?
Affected Groups: Will this feel fair, transparent, and beneficial to them?
Functional Teams: Do we have the capability and resources to sustain this?
External Stakeholders: Unions, regulators, partners, customers—will they support or oppose?

7.9 Common AI Issues Across Disciplines and How to Fix Them

7.9.1 Issue 1: Overcomplicated Solutions

AI Tendency: Creates comprehensive but unimplementable recommendations

Example: Tourism & Hospitality

AI might suggest: 20-touchpoint guest journey map with personalisation at every stage
Better: Focus on the 5 moments that most influence guest satisfaction and rebooking

Your Direction: “Start with the 80/20 rule—what 20% will address 80% of situations?”

7.9.2 Issue 2: Ignoring Compliance and Constraints

AI Tendency: Focuses on best practices without considering legal, technical, or operational realities

Example: Economics

AI might write: "Reduce interest rates to stimulate growth"
Better: "Evaluate rate adjustments considering inflation expectations, exchange rate effects, household debt levels, and transmission mechanism lags"

Your Direction: “What are our legal/technical/operational obligations and constraints?”

7.9.3 Issue 3: One-Size-Fits-All Recommendations

AI Tendency: Provides generic advice without organisational context

Example: Finance

AI might suggest: "Apply standard financial ratios for all investment decisions"
Better: "Apply ratios adjusted for industry, company stage, and investment type"

Your Direction: “How should this be adapted for our specific context, constraints, and different situations?”

7.10 Your Critique Conversation Templates

7.10.1 Template 1: Requesting Simplification

“This solution looks more complex than what we can realistically implement. I work in a 200-employee manufacturing company with a small HR team. Can you give me a practical version that focuses on the essentials and doesn’t require expensive software or additional staff?”

7.10.2 Template 2: Checking Legal Compliance

“I need to ensure this recommendation complies with Australian employment law. What specific legislation or legal requirements should I consider? Are there any potential discrimination risks or privacy concerns I need to address?”

7.10.3 Template 3: Testing Organisational Fit

“Before I present this to senior management, I need to understand how this would work in our context. We have a unionised workforce, high employee turnover in customer service roles, and managers who are time-poor. How should I adapt this recommendation for our specific situation?”

7.10.4 Template 4: Anticipating Resistance

“What resistance should I expect if I implement this recommendation? Walk me through the likely concerns from employees, middle managers, and senior leadership. How can I address these concerns proactively?”

7.11 Your HR Professional Documentation

After each AI interaction, document your critique process:

Template:

# AI Interaction #[number]
**My Request**: [What HR problem I asked AI to solve]
**AI's First Response**: [Brief summary of the recommendation]
**My Critique**: [What I questioned and requested to improve]
**Final Solution**: [What we ended up with after iteration]
**Implementation Considerations**: [What I still need to check/adapt for my workplace]
**What I Learned**: [Key insight for future AI interactions]

Example:

# AI Interaction #3
**My Request**: Create a new employee onboarding checklist for a 50-employee tech company
**AI's First Response**: Comprehensive 30-day checklist with daily activities and multiple stakeholder meetings
**My Critique**: Too intensive for our small team, assumes dedicated onboarding coordinator
**Final Solution**: 5-day essential checklist with weekly follow-ups for first month
**Implementation Considerations**: Need to check which IT systems can be automated, get manager buy-in for time allocation
**What I Learned**: Always ask for scalable solutions that don't require additional headcount

7.12 Red Flag Checklist for Any Recommendation

Before accepting any AI-generated recommendation, ask:

Can I explain this solution clearly to key stakeholders in a few minutes?
Does this comply with relevant laws, regulations, and company policies?
Do we have the resources and capability to implement this?
How will affected stakeholders experience this—is it fair and transparent?
What are the risks if this implementation goes poorly?
Have I considered how different groups or departments might be affected?
Is there a simpler version that would achieve 80% of the benefits?

7.13 Practice: Critique This AI Response

AI Generated Recommendation:

To improve employee engagement, implement a comprehensive recognition program including:
1. Monthly peer-to-peer recognition awards with monetary prizes
2. Quarterly manager-nominated excellence awards with public ceremonies
3. Annual employee of the year with significant financial bonus
4. Real-time digital recognition platform with social features
5. Team-based performance incentives with quarterly payouts

What’s Wrong? (Think before checking the answer)

Issues to Critique

Overcomplicated: Five different recognition systems is confusing and administratively heavy
No Legal Consideration: No mention of tax implications, fairness, or potential discrimination
Assumes Budget: Significant financial costs without ROI justification
One-Size-Fits-All: Doesn’t consider different employee preferences (public vs private recognition)
No Context: Doesn’t consider company culture, size, or existing systems
Implementation Gap: No thought to how managers will administer this fairly

Your Response: “This is too complex and expensive for our needs. Can you suggest a simpler recognition approach that focuses on meaningful acknowledgment rather than monetary rewards, and that our managers can implement without significant administrative burden?”

7.14 Teaching Students to Critique AI

7.14.1 Classroom Exercise: The AI Consultant Swap

Setup: Divide students into small groups. Give each group a different business problem relevant to their discipline (e.g., HR: turnover; Finance: investment strategy; Supply Chain: supplier consolidation; Marketing: campaign strategy; Management: change management).

Task: 1. Each group uses AI to generate a solution to their problem 2. Groups swap their AI-generated solutions with another group 3. Each group must critique the other group’s AI solution using the 5-step framework 4. Groups present both the original AI solution and their critique to the class

Learning Outcome: Students experience both generating AI solutions and critically evaluating them, understanding that the real value lies in the critique process.

7.14.2 Assessment Idea: AI Solution Critique

Assignment Requirements (adapted by discipline): 1. Choose a business challenge relevant to your workplace or placement organisation 2. Use AI to generate three different approaches to solving this challenge 3. Critique each approach using the 5-step framework 4. Recommend which approach (or combination) is most suitable for your specific context 5. Justify your recommendation with reference to discipline-specific theory, compliance requirements, and organisational considerations

What You’re Assessing: - Critical thinking about AI-generated solutions - Understanding of organisational context and constraints - Legal/technical/operational and ethical awareness - Ability to translate theory into practical recommendations - Professional judgment in evaluating AI outputs

7.15 Connecting to the VET Framework

The five-step critique framework in this chapter teaches students to evaluate AI output in context, checking for practicality, compliance, assumptions, and stakeholder impact. There is a complementary framework that works at a more fundamental level.

The VET framework, from Conversation, Not Delegation, asks three questions before acting on any AI output:

Verify: Can I find this independently?
Explain: Can I explain this in my own words?
Test: Does this hold up under scrutiny?

The two frameworks reinforce each other. VET catches the foundational failures: fabricated claims, shallow understanding, fragile reasoning. The business critique framework catches the contextual failures: solutions that are technically correct but wrong for your organisation, your constraints, your stakeholders.

Teaching both gives students a complete critical toolkit: VET for “is this true and do I understand it?” and the five-step framework for “is this right for my situation?” For a deeper treatment of VET and the cognitive traps that undermine critical evaluation, see Conversation, Not Delegation.

Try this (2 minutes)

Paste a piece of your own teaching material into AI and ask “How is this?” Note the response — it will almost certainly be positive. Now ask “Identify the three weakest aspects of this and explain why each one could be improved.” Compare the two responses. The difference reveals how much the AI was telling you what you wanted to hear the first time. Teach your students to do the same: always ask for specific criticism, not general impressions.

Three questions before you act

When AI produces something that will reach students or colleagues, pause for thirty seconds and ask: Can I independently confirm the key claims? Could I explain the reasoning in my own words to a sceptical colleague? Would I be comfortable defending this in a course review meeting? If any answer is no, the output needs more work — yours, not the AI’s.

7.16 Why This Matters for Professional Careers

In the next five years, professionals in every business discipline will work alongside AI tools. The ones who thrive won’t be those who can generate the most impressive AI outputs—they’ll be those who can skillfully evaluate, adapt, and improve AI recommendations.

Critical thinking about AI is becoming a core professional competency across all disciplines.

Employers will increasingly ask: - “How do you use AI in your work?” - “How do you ensure AI recommendations are appropriate for our organisation?” - “Can you give an example of when you identified problems with an AI-generated solution?”

Students who master the critique framework will have compelling answers to these questions. They’ll demonstrate that they’re not just AI users—they’re AI-savvy professionals who can leverage technology while maintaining professional judgment and ethical standards.

7.17 Your Action Step

Before moving to the next chapter, practice the critique framework:

Choose a business challenge in your discipline that you’re currently facing or teaching about
Ask an AI tool for a recommendation or solution
Apply the 5-step critique framework to identify issues and improvements
Iterate with the AI until you have a solution you’d actually implement
Document your process using the template provided

This hands-on experience will help you teach students to be thoughtful, critical users of AI rather than passive consumers of AI-generated content.

# The AI Critique Toolkit - Becoming a Smart Business Professional > The most dangerous AI output is not the one that is obviously wrong. It is the one that sounds exactly right, and nobody checks. ## Why Business Students Need Critique Skills Imagine you just hired a brilliant but inexperienced analyst or consultant. They work fast, have lots of ideas, but sometimes: - Overcomplicate simple solutions - Miss important legal, technical, or operational considerations - Write recommendations that sound good but have practical flaws - Make assumptions about your workplace context, capacity, or constraints **This is exactly how AI behaves.** Your job as a business professional is to review, question, and improve AI-generated advice before it impacts real people and real organisations. The difference between good business professionals and great ones isn't whether they use AI tools---it's how critically they evaluate AI outputs. In professional work across all disciplines, bad advice can lead to legal challenges, damaged relationships, failed initiatives, or strategic missteps. ## The Business Impact of Uncritical AI Acceptance **Scenario Examples Across Disciplines**: ::: {.callout-tip title="Example: Management"} Your AI generates a change management plan. You follow it without stakeholder testing. Outcomes: - Key influencers weren't engaged early enough - Communication messaging doesn't resonate with your culture - Implementation timeline is unrealistic given other priorities - Change initiative stalls after initial enthusiasm ::: **Lesson**: Always critique AI-generated recommendations thoroughly before implementation, regardless of discipline. --- ## Your 5-Step Business Critique Framework This framework applies across all business disciplines: 1. **Comprehension Check** - "Do I understand this completely?" 2. **Simplicity Check** - "Is this practical for my workplace/situation?" 3. **Legal, Technical & Risk Check** - "What are the legal, technical, ethical, and operational implications?" 4. **Context Check** - "What assumptions is AI making about my organisation?" 5. **Stakeholder Check** - "How will different groups react to this?" --- ## Step 1: Comprehension Check - "Do I understand this?" **Red Flags**: - HR jargon that sounds impressive but is unclear - Policy language that could be interpreted multiple ways - No clear explanation of why this approach is recommended **Your Response**: - "Can you explain this in plain English that a line manager could understand?" - "What specific problem does this solve and how does it solve it?" - "Break this down into step-by-step actions that need to be taken" **Example**: ``` # ❌ AI gives you this: "We recommend implementing a holistic performance ecosystem leveraging" "synergistic feedback loops and agile recalibration mechanisms." # ✅ Ask for this instead: "We recommend creating a performance management system where employees get feedback from multiple sources throughout the year, not just once annually. This includes regular check-ins with managers, peer feedback on projects, and quarterly progress reviews against clear goals." ``` --- ## Step 2: Simplicity Check - "Is this practical for my workplace?" **Red Flags**: - Solution requires resources you don't have (extra HR staff, expensive software) - Process is more complex than your current problem - Assumes perfect implementation with no room for human error **Your Response**: - "Give me a version that works with a 2-person HR team and limited budget" - "What's the minimum viable version of this solution?" - "Show me how to implement this step by step over 6 months, not all at once" **HR Example**: AI suggests a sophisticated 360-degree feedback system with custom software, external facilitators, and detailed analytics. You ask for a simpler approach and get a practical solution using existing tools like Google Forms and manager training sessions. --- ## Step 3: Legal, Technical & Risk Check - "What are the implications?" **Red Flags** (vary by discipline): - No mention of compliance, legal, technical, or regulatory considerations - Recommendations that could harm certain groups (employees, customers, stakeholders) - Risk implications that aren't addressed - No consideration of organisational constraints or industry requirements **Your Response** (discipline-specific examples): ::: {.callout-tip title="Example: Supply Chain"} - "What are the operational and financial risks?" - "How does this affect compliance and sustainability?" - "What supply chain resilience issues exist?" ::: **Critical Questions to Always Ask** (across all disciplines): - What could go wrong and what's the exposure? - Are we meeting our compliance and governance obligations? - What organisational constraints or capabilities might we lack? - Who might be negatively affected and how? --- ## Step 4: Context Check - "What assumptions is AI making about my organisation?" **Red Flags**: - AI assumes you have a sophisticated HR information system - Assumes your managers are all skilled in difficult conversations - Assumes your workplace culture is collaborative and high-trust - Assumes employees are digitally literate and open to change **Your Response**: - "What assumptions are you making about our current systems and processes?" - "How would this work in a workplace with high employee turnover?" - "What if our managers resist this change?" - "How does this account for our hybrid/remote/in-person work arrangement?" **Real Example**: AI recommends an employee engagement app with daily mood tracking and gamified recognition. You realise this assumes your employees are comfortable sharing personal data and that you have the technical infrastructure to support it. You ask for alternatives that work with your existing communication channels. --- ## Step 5: Stakeholder Check - "How will different groups react to this?" **Red Flags**: - Solution only considers management perspective - No thought to how employees will perceive or experience the change - Doesn't address how different departments might be affected differently - No consideration of change management requirements **Your Response**: - "Walk me through how this would feel from an employee's perspective" - "How might different departments (sales, operations, finance) experience this differently?" - "What resistance should we anticipate and how can we address it?" - "What communication and training would be needed for successful implementation?" **Stakeholder Mapping** (adjust by discipline): Always consider: - **Senior Leadership**: Will they see the business value and strategic alignment? - **Operational Staff/Managers**: Do they have skills, capacity, and buy-in to implement? - **Affected Groups**: Will this feel fair, transparent, and beneficial to them? - **Functional Teams**: Do we have the capability and resources to sustain this? - **External Stakeholders**: Unions, regulators, partners, customers---will they support or oppose? --- ## Common AI Issues Across Disciplines and How to Fix Them ### Issue 1: Overcomplicated Solutions **AI Tendency**: Creates comprehensive but unimplementable recommendations ::: {.callout-tip title="Example: Tourism & Hospitality"} ``` AI might suggest: 20-touchpoint guest journey map with personalisation at every stage Better: Focus on the 5 moments that most influence guest satisfaction and rebooking ``` ::: **Your Direction**: "Start with the 80/20 rule---what 20% will address 80% of situations?" ### Issue 2: Ignoring Compliance and Constraints **AI Tendency**: Focuses on best practices without considering legal, technical, or operational realities ::: {.callout-tip title="Example: Economics"} ``` AI might write: "Reduce interest rates to stimulate growth" Better: "Evaluate rate adjustments considering inflation expectations, exchange rate effects, household debt levels, and transmission mechanism lags" ``` ::: **Your Direction**: "What are our legal/technical/operational obligations and constraints?" ### Issue 3: One-Size-Fits-All Recommendations **AI Tendency**: Provides generic advice without organisational context ::: {.callout-tip title="Example: Finance"} ``` AI might suggest: "Apply standard financial ratios for all investment decisions" Better: "Apply ratios adjusted for industry, company stage, and investment type" ``` ::: **Your Direction**: "How should this be adapted for our specific context, constraints, and different situations?" --- ## Your Critique Conversation Templates ### Template 1: Requesting Simplification *"This solution looks more complex than what we can realistically implement. I work in a 200-employee manufacturing company with a small HR team. Can you give me a practical version that focuses on the essentials and doesn't require expensive software or additional staff?"* ### Template 2: Checking Legal Compliance *"I need to ensure this recommendation complies with Australian employment law. What specific legislation or legal requirements should I consider? Are there any potential discrimination risks or privacy concerns I need to address?"* ### Template 3: Testing Organisational Fit *"Before I present this to senior management, I need to understand how this would work in our context. We have a unionised workforce, high employee turnover in customer service roles, and managers who are time-poor. How should I adapt this recommendation for our specific situation?"* ### Template 4: Anticipating Resistance *"What resistance should I expect if I implement this recommendation? Walk me through the likely concerns from employees, middle managers, and senior leadership. How can I address these concerns proactively?"* --- ## Your HR Professional Documentation After each AI interaction, document your critique process: **Template**: ``` # AI Interaction #[number] **My Request**: [What HR problem I asked AI to solve] **AI's First Response**: [Brief summary of the recommendation] **My Critique**: [What I questioned and requested to improve] **Final Solution**: [What we ended up with after iteration] **Implementation Considerations**: [What I still need to check/adapt for my workplace] **What I Learned**: [Key insight for future AI interactions] ``` **Example**: ``` # AI Interaction #3 **My Request**: Create a new employee onboarding checklist for a 50-employee tech company **AI's First Response**: Comprehensive 30-day checklist with daily activities and multiple stakeholder meetings **My Critique**: Too intensive for our small team, assumes dedicated onboarding coordinator **Final Solution**: 5-day essential checklist with weekly follow-ups for first month **Implementation Considerations**: Need to check which IT systems can be automated, get manager buy-in for time allocation **What I Learned**: Always ask for scalable solutions that don't require additional headcount ``` --- ## Red Flag Checklist for Any Recommendation Before accepting any AI-generated recommendation, ask: - **Can I explain this solution clearly to key stakeholders in a few minutes?** - **Does this comply with relevant laws, regulations, and company policies?** - **Do we have the resources and capability to implement this?** - **How will affected stakeholders experience this---is it fair and transparent?** - **What are the risks if this implementation goes poorly?** - **Have I considered how different groups or departments might be affected?** - **Is there a simpler version that would achieve 80% of the benefits?** --- ## Practice: Critique This AI Response **AI Generated Recommendation**: ``` To improve employee engagement, implement a comprehensive recognition program including: 1. Monthly peer-to-peer recognition awards with monetary prizes 2. Quarterly manager-nominated excellence awards with public ceremonies 3. Annual employee of the year with significant financial bonus 4. Real-time digital recognition platform with social features 5. Team-based performance incentives with quarterly payouts ``` **What's Wrong?** (Think before checking the answer) <details> <summary>Issues to Critique</summary> 1. **Overcomplicated**: Five different recognition systems is confusing and administratively heavy 2. **No Legal Consideration**: No mention of tax implications, fairness, or potential discrimination 3. **Assumes Budget**: Significant financial costs without ROI justification 4. **One-Size-Fits-All**: Doesn't consider different employee preferences (public vs private recognition) 5. **No Context**: Doesn't consider company culture, size, or existing systems 6. **Implementation Gap**: No thought to how managers will administer this fairly **Your Response**: "This is too complex and expensive for our needs. Can you suggest a simpler recognition approach that focuses on meaningful acknowledgment rather than monetary rewards, and that our managers can implement without significant administrative burden?" </details> --- ## Teaching Students to Critique AI ### Classroom Exercise: The AI Consultant Swap **Setup**: Divide students into small groups. Give each group a different business problem relevant to their discipline (e.g., HR: turnover; Finance: investment strategy; Supply Chain: supplier consolidation; Marketing: campaign strategy; Management: change management). **Task**: 1. Each group uses AI to generate a solution to their problem 2. Groups swap their AI-generated solutions with another group 3. Each group must critique the other group's AI solution using the 5-step framework 4. Groups present both the original AI solution and their critique to the class **Learning Outcome**: Students experience both generating AI solutions and critically evaluating them, understanding that the real value lies in the critique process. ### Assessment Idea: AI Solution Critique **Assignment Requirements** (adapted by discipline): 1. Choose a business challenge relevant to your workplace or placement organisation 2. Use AI to generate three different approaches to solving this challenge 3. Critique each approach using the 5-step framework 4. Recommend which approach (or combination) is most suitable for your specific context 5. Justify your recommendation with reference to discipline-specific theory, compliance requirements, and organisational considerations **What You're Assessing**: - Critical thinking about AI-generated solutions - Understanding of organisational context and constraints - Legal/technical/operational and ethical awareness - Ability to translate theory into practical recommendations - Professional judgment in evaluating AI outputs --- ## Connecting to the VET Framework The five-step critique framework in this chapter teaches students to evaluate AI output in context, checking for practicality, compliance, assumptions, and stakeholder impact. There is a complementary framework that works at a more fundamental level. The **VET framework**, from *Conversation, Not Delegation*, asks three questions before acting on any AI output: - **Verify:** Can I find this independently? - **Explain:** Can I explain this in my own words? - **Test:** Does this hold up under scrutiny? The two frameworks reinforce each other. VET catches the foundational failures: fabricated claims, shallow understanding, fragile reasoning. The business critique framework catches the contextual failures: solutions that are technically correct but wrong for your organisation, your constraints, your stakeholders. Teaching both gives students a complete critical toolkit: VET for "is this true and do I understand it?" and the five-step framework for "is this right for my situation?" For a deeper treatment of VET and the cognitive traps that undermine critical evaluation, see *Conversation, Not Delegation*. ::: {.callout-tip title="Try this (2 minutes)"} Paste a piece of your own teaching material into AI and ask "How is this?" Note the response — it will almost certainly be positive. Now ask "Identify the three weakest aspects of this and explain why each one could be improved." Compare the two responses. The difference reveals how much the AI was telling you what you wanted to hear the first time. Teach your students to do the same: always ask for specific criticism, not general impressions. ::: ::: {.callout-important title="Three questions before you act"} When AI produces something that will reach students or colleagues, pause for thirty seconds and ask: Can I independently confirm the key claims? Could I explain the reasoning in my own words to a sceptical colleague? Would I be comfortable defending this in a course review meeting? If any answer is no, the output needs more work — yours, not the AI's. ::: --- ## Why This Matters for Professional Careers In the next five years, professionals in every business discipline will work alongside AI tools. The ones who thrive won't be those who can generate the most impressive AI outputs---they'll be those who can skillfully evaluate, adapt, and improve AI recommendations. **Critical thinking about AI is becoming a core professional competency across all disciplines.** Employers will increasingly ask: - "How do you use AI in your work?" - "How do you ensure AI recommendations are appropriate for our organisation?" - "Can you give an example of when you identified problems with an AI-generated solution?" Students who master the critique framework will have compelling answers to these questions. They'll demonstrate that they're not just AI users---they're AI-savvy professionals who can leverage technology while maintaining professional judgment and ethical standards. --- ## Your Action Step Before moving to the next chapter, practice the critique framework: 1. **Choose a business challenge** in your discipline that you're currently facing or teaching about 2. **Ask an AI tool** for a recommendation or solution 3. **Apply the 5-step critique framework** to identify issues and improvements 4. **Iterate with the AI** until you have a solution you'd actually implement 5. **Document your process** using the template provided This hands-on experience will help you teach students to be thoughtful, critical users of AI rather than passive consumers of AI-generated content. ---