Essential introduction to Generative AI concepts, capabilities, limitations, and ethical considerations for academic staff
Understanding Generative AI
What is Generative AI?
Generative Artificial Intelligence (GenAI) refers to AI systems that can create new content—text, images, code, audio, or other media—based on patterns learned from training data. Unlike traditional AI that classifies or predicts, GenAI generates original content.
🔍 Key Characteristics
Pattern Recognition: Learns from vast amounts of existing content
Content Creation: Generates new, original outputs based on prompts
Context Awareness: Understands and responds to specific contexts and instructions
Iterative Improvement: Can refine outputs based on feedback
Capabilities & Limitations
GenAI Capabilities and Limitations Comparison
What GenAI Can Do
What GenAI Cannot Do
Generate educational content quickly
Replace human judgment and expertise
Adapt content for different audiences
Ensure 100% accuracy without verification
Suggest multiple approaches and formats
Understand your specific student contexts
Help with accessibility considerations
Replace the need for accessibility expertise
Provide creative examples and analogies
Guarantee cultural sensitivity without guidance
AI in Academic Contexts
In educational settings, GenAI can be particularly powerful for:
📚 Content Creation
Lecture notes and summaries
Assessment questions
Reading materials at different levels
Visual descriptions for accessibility
🎯 Personalization
Adapting content for different learning styles
Language simplification for EAL students
Multiple format versions (text, lists, diagrams)
Scaffolding for different ability levels
⚡ Efficiency
Quick first drafts to iterate upon
Feedback templates for assessments
Email responses and communications
Administrative task automation
⚠️ Important: GenAI is a co-pilot, not a replacement. Your expertise, judgment, and understanding of your students' needs remain essential for effective teaching and learning.
Ethics & Responsible Use
Understanding Bias in AI
GenAI systems can perpetuate and amplify biases present in their training data. These biases can affect:
Cultural representation: Examples may default to dominant cultural perspectives
Language patterns: May favor certain linguistic styles or registers
Demographic assumptions: Could make assumptions about student backgrounds
Disciplinary perspectives: May reflect established academic hierarchies and viewpoints
🔍 Bias Detection Prompt
Review this content for potential bias:
[PASTE YOUR AI-GENERATED CONTENT HERE]
Please analyze for:
1. Cultural assumptions or perspectives that may not be universal
2. Language that may exclude or disadvantage certain groups
3. Examples that may not represent diverse backgrounds
4. Implicit assumptions about prior knowledge or experience
5. Suggestions for making this content more inclusive
Provide specific recommendations for improvement.
Transparency and Academic Integrity
When using GenAI in educational contexts, transparency is crucial:
Transparency Statement Template:
"Parts of this [lecture/material/assessment] were created with assistance from generative AI tools. Specifically, AI was used to:
- [List specific uses, e.g., "generate initial examples"]
- [e.g., "suggest alternative explanations"]
- [e.g., "create accessibility descriptions"]
All AI-generated content has been reviewed, fact-checked, and adapted to ensure accuracy and appropriateness for our learning context."
Prompt Engineering Frameworks
Effective prompting is the key to getting useful outputs from GenAI. Here are five proven frameworks you can use immediately:
1. P-T-C-F Framework
Best for: Content creation, lesson planning, clear task-oriented requests
Persona: Who should the AI be? (e.g., "You are an experienced biology educator")
Task: What exactly do you want? (e.g., "Create a lesson plan on...")
Context: Background information (e.g., "For first-year students who struggle with...")
Format: How should the output look? (e.g., "As a structured outline with activities")
📋 P-T-C-F Template
Persona: You are [role/expertise]
Task: [Specific action you want]
Context:
- Audience: [who are your students/colleagues]
- Background: [relevant information]
- Constraints: [any limitations or requirements]
Format: Please provide this as [specific format/structure]
Example - Creating Accessible Content:Persona: You are an inclusive education specialist with expertise in Universal Design for Learning.
Task: Create an explanation of photosynthesis that can be understood by students with varying learning needs.
Context:
- Audience: First-year undergraduate students in an introductory biology course
- Background: Mixed abilities, some students have learning differences, English as additional language learners
- Constraints: Must be scientifically accurate but accessible
Format: Provide the explanation in three formats: 1) Simple text summary, 2) Step-by-step process with visual cues, 3) Analogy-based explanation
2. Chain-of-Thought
Best for: Complex processes, problem-solving, breaking down difficult concepts
Ask the AI to think through problems step-by-step, showing its reasoning process.
Chain-of-Thought Example:
"I need to explain how enzymes work to students who are struggling with the concept.
Think through this step-by-step:
1. First, identify what makes enzymes difficult to understand
2. Then, consider what analogies or examples might help
3. Think about potential misconceptions students might have
4. Finally, design an explanation that addresses these challenges
Please show your thinking process for each step, then provide your final explanation."
3. Few-Shot Learning
Best for: Maintaining consistency, following specific patterns, creating content in your style
Provide 2-3 examples of what you want, then ask for more in the same style.
Few-Shot Example - Accessible Question Writing:
"I want to create multiple-choice questions that are accessible to all students. Here are examples of my style:
Example 1:
Question: What is the primary function of chloroplasts in plant cells?
A) To store water and maintain cell pressure
B) To convert light energy into chemical energy through photosynthesis
C) To break down waste materials in the cell
D) To provide structural support to the cell wall
Example 2:
Question: Which process allows nutrients to move from the soil into plant roots?
A) Active transport using energy from the cell
B) Photosynthesis converting sunlight to energy
C) Respiration breaking down stored sugars
D) Transpiration moving water through the plant
Now create 3 more questions about cellular respiration following the same style and accessibility principles."
4. Role-Based Prompting
Best for: Expert-level content, authoritative responses, specialized knowledge
Have the AI adopt a specific expert role with defined background and expertise.
Role-Based Example:
"You are Dr. Sarah Chen, a learning disabilities specialist with 15 years of experience in higher education accessibility. You have particular expertise in supporting students with dyslexia, ADHD, and autism in STEM subjects.
A colleague asks you: 'How can I make my biochemistry lectures more accessible for students with attention difficulties?'
Respond as Dr. Chen would, drawing on your specific experience and providing practical, evidence-based recommendations."
5. CLEAR Framework
Best for: Comprehensive lesson planning, detailed educational content
Context: Background and setting
Length: How much content you need
Examples: Include specific examples
Audience: Who is this for
Role: What role should AI adopt
CLEAR Template:Context: [Course, topic, learning objectives]
Length: [Duration/word count needed]
Examples: [Include/reference specific examples needed]
Audience: [Student demographics, level, needs]
Role: [What expertise should AI demonstrate]
🚀 Ready for Advanced Techniques?
These frameworks are just the beginning! For deeper exploration of prompt engineering, including advanced techniques like prompt chaining, conditional logic, and specialized academic applications, check out our Advanced Prompting Workshop.
Inclusive Design with AI
When we design educational content and experiences, we can either build barriers or remove them. GenAI gives us powerful tools to create inclusive learning environments from the start.
Universal Design for Learning Meets AI
Universal Design for Learning (UDL) is built on three core principles that align perfectly with GenAI capabilities:
UDL Principles and GenAI Applications
UDL Principle
How GenAI Helps
Example Applications
Multiple Means of Engagement The "why" of learning
Generate diverse examples, cultural contexts, and personal relevance connections
Create examples from different cultural backgrounds, career applications, personal interest connections
Multiple Means of Representation The "what" of learning
Transform content into multiple formats, languages, and complexity levels
Convert text to lists, create analogies, simplify language, add visual descriptions
Multiple Means of Action/Expression The "how" of learning
Design varied assessment formats and provide scaffolding options
Create alternatives to essays, design rubrics, provide templates and frameworks
🎯 UDL-Focused Prompt Template
Create [content/activity] that follows Universal Design for Learning principles:
Topic: [Your subject matter]
Audience: [Student demographics and needs]
Multiple Engagement: Include diverse examples and cultural connections
Multiple Representations: Provide content in [visual/text/audio/etc.] formats
Multiple Expressions: Offer [variety of ways students can demonstrate learning]
Accessibility Requirements:
- Language level: [appropriate for audience]
- Visual descriptions: [for any graphics/diagrams]
- Cultural sensitivity: [acknowledge diverse backgrounds]
- Cognitive load: [appropriate complexity and scaffolding]
Web Content Accessibility Guidelines (WCAG)
When creating digital educational content, we follow WCAG 2.1 AA standards. GenAI can help implement these principles:
🔤 Perceivable
Information must be presentable in ways users can perceive
Alternative text for images
Captions for videos
High color contrast
Resizable text
🖱️ Operable
Interface components must be operable by all users
Keyboard navigation
No seizure-inducing flashing
Sufficient time for tasks
Clear navigation
🧠 Understandable
Information and UI operation must be understandable
Clear, simple language
Predictable functionality
Input assistance and error prevention
Consistent terminology
💪 Robust
Content must be robust enough for assistive technologies
Valid HTML markup
Compatible with screen readers
Semantic structure
ARIA labels where needed
📖 WCAG Evolution:
WCAG 2.0 (2008): Established core accessibility principles
WCAG 2.1 (2018): Added mobile accessibility and cognitive support - current standard
WCAG 3.0 (in development): Will include better testing methods and broader disability coverage
For now, focus on WCAG 2.1 AA compliance, which is the legal standard in the UK under the Equality Act 2010.
Designing for All Learners
Inclusive Prompting Strategies
When creating prompts, build inclusion into your requests:
Inclusive Assessment Design Prompt:
Create an assessment for [topic] that provides multiple ways for students to demonstrate their learning:
Learning Outcomes: [List specific outcomes]
Student Needs: Consider students who may have:
- Different language backgrounds (EAL learners)
- Learning differences (dyslexia, ADHD, autism)
- Varied cultural experiences and perspectives
- Different confidence levels with written expression
Assessment Options: Provide at least 3 different formats:
1. [Traditional format, e.g., written report]
2. [Visual/multimedia option, e.g., infographic + presentation]
3. [Interactive/practical option, e.g., demonstration + reflection]
Success Criteria: Ensure each option allows students to show the same depth of understanding through different means.
Support Materials: Include templates, rubrics, and examples for each format.
Language-Inclusive Content Creation:
Transform this content to be more accessible for English as Additional Language (EAL) learners:
[PASTE YOUR CONTENT]
Requirements:
- Simplify complex sentences without losing meaning
- Define technical terms clearly
- Provide cultural context for UK-specific references
- Use active voice where possible
- Include transition words and clear signposting
- Add examples that connect to diverse cultural experiences
- Suggest visual aids or diagrams to support understanding
Output: Provide the revised content plus a glossary of key terms.
🌟 Deep Dive into Inclusive Teaching:
Ready to explore specific strategies for accessibility audits, EAL support, and inclusive assessment design? Check out our comprehensive Inclusive Teaching with GenAI Workshop for practical tools and detailed guidance.
The Anticipatory Duty
💡 Key Principle: Design for the Margins
When we design for students who need the most support—those with learning differences, language barriers, or other challenges—we create solutions that benefit everyone.
Clear structure helps all students navigate content
Multiple formats give everyone options to play to their strengths
Simple language improves comprehension for all readers
Cultural examples create a more welcoming learning environment
Result: Instead of retrofitting accessibility, we build inclusive experiences from the start.