Getting Our Feet Wet: Navigating the AI Tsunami in Higher Education
Rachel Koblic
4 December 2025 | Dublin, Ireland
Who Am I?
AI-generated headshot via BetterPic.io
Learning Architect & Product Strategist
I've spent 15+ years designing learning, most recently as Chief Learning Officer at Matter and Space building AI-first education alongside Paul LeBlanc and George Siemens.
AI-generated image via Midjourney
AI-generated image via Midjourney
Agenda
01
AI State of the Union
Where we are today in the AI revolution
02
AI in Higher Education
Transforming teaching, learning, and quality assurance
03
AI Day-to-Day
Practical tools and strategies for immediate impact
AI State of the Union
Where are we today?
Timeline
1
1956: Dartmouth Conference
Birth of artificial intelligence as a field: Can machines think and learn like humans do?
2
1960s-2010s: Decades of Evolution
Experimentation in research labs, early breakthroughs, and AI winters
3
2012: Deep Learning Revolution
Neural networks transform what's possible: the more data you give them the better they get
4
2022: ChatGPT's Launch
AI enters mainstream consciousness and acceleration takes off
ChatGPT's Explosive Growth
100 Million Users in 2 Months
Fastest-growing consumer application in history
Global Usage Surge
58%
Use AI for Work
KPMG & University of Melbourne study of 48,000 people across 47 countries
Explosion in Research
40% of all CS Publications
Up from 22% a decade ago. Academic publications and research papers in AI have grown exponentially.
And Investment
US $109 Billion
Investment in AI continues to surge globally.
AI's Performance Trajectory
Matching or Exceeding Human Performance
AI systems now match or exceed human performance across multiple benchmark tasks
  • Image classification
  • Reading comprehension
  • Multitask language understanding
  • PhD-level science questions
  • Visual reasoning
  • English langage understanding
  • Competition-level math
Ethan Mollick's "Otter on a plane using wifi"
The Challenges
Hallucinations
  • Best models <1% error on standard tasks
  • Reasoning models 30-50% on multi-step thinking
  • Sounds authoritative even when completely wrong
Bias & Representation
  • Women linked to "home/family" 4x more than men
  • Most training data is English/U.S.-centric
  • Western cultural values dominate all models
Sustainability
  • ChatGPT queries use 10x energy vs. Google search
  • AI data centers: 2 TWh (2017) → 40+ TWh (2023) and set to double by 2030
  • 30% of AI training power wasted on inefficiency
Reasons for Optimism: Healthcare
AI is revolutionizing diagnostics, drug discovery, and patient care with unprecedented accuracy and speed
Reasons for Optimism: Science & Climate
From protein folding to climate modeling, AI accelerates scientific breakthroughs that address humanity's greatest challenges
Reasons for Optimism: Education
Personalized tutoring, adaptive learning paths, and instant feedback democratize quality education at scale
Amplifying Human Potential
AI in Higher Education & QA
How does all of this impact us?
My Experience at Matter and Space
Dynamically Generated Learning
Almost entirely personalized experiences powered by:
  • Knowledge graph: Interconnected concepts and relationships
  • AI agents: Intelligent content generation and adaptation
  • Learner profile: Individualized pathways and pacing
Five Questions for QA in the Age of AI
1
Learning Outcomes
What do learners need to learn?
2
Teaching & Learning
How will teaching & learning change?
3
Assessment Evidence
How do we assess progress and proficiency?
4
Objects of Review
What new materials will QA teams evaluate?
5
Capabilities & Capacity
What do you need?
What skills do learners need?
Demand for GenAI is Growing
Mentions of GenAI in job postings grew 4× in one year between 2023 and 2024.
Source: Lightcast
Traditional vs. AI-Era Competencies
"Soft" Skills More Important Than Ever
Ask the Right Questions
Frame problems effectively for AI collaboration
Critique Model Outputs
Evaluate AI-generated content with discernment
Detect Hallucinations & Bias
Identify when AI produces inaccurate or biased results
Handle Ambiguity
Navigate uncertain and complex situations
Apply Ethical Reasoning
Make principled decisions about AI use
Document Decision-Making
Create clear audit trails for AI-assisted work
How will teaching & learning change?
Faculty Roles Evolve
  • Augmented, not replaced
  • New pedagogical competencies required
  • Credentials and professional development must adapt
AI-Powered Tutors
  • Always-available, personalized tutoring
  • Potential for Bloom's two sigma effect at scale
  • Challenges traditional credit hour definitions
Dynamic Curriculum Delivery
  • Content generated on-the-fly by AI agents
  • Guided by system prompts and learner profiles
  • Raises questions about what we review…
Everyone is "Cheating"
How do we assess learning?
Before AI
Product = Evidence
Final work demonstrated learning
With AI
Product ≠ Evidence
Final work may be AI-generated
Solution
Process + Product
Document the journey & decisions

Capture the Full Story: Prompts, drafts, critiques, and decision-making processes reveal authentic learning
What exactly will QA need to review?
The Familiar, Evolved
Faculty CVs
Now include AI pedagogical competency
Assessment Plans
Incorporate process evidence & decision logs
Course Materials
Document AI integration & boundaries
The Entirely New
Agent Documentation
How AI systems operate and make decisions
Evaluation Logs
Records of AI output verification
System Prompts
Instructions guiding AI behavior
Calibration Data
Evidence of AI alignment with learning goals
Graphs are the New Syllabi
Traditional Syllabus
  • Linear, sequential content
  • Fixed learning path
  • One-size-fits-all structure
  • Static document
Knowledge Graph
  • Interconnected concepts
  • Multiple pathways
  • Personalized journeys
  • Dynamic adaptation
What do YOU need?
Essential Resources for Higher Ed Teams
  • Training: Understanding AI capabilities and limitations
  • Frameworks: New quality criteria and rubrics
  • Tools: Technology to review AI-generated content
  • Time: Space to develop new competencies
  • Community: Peer networks for shared learning
AI Day-to-Day
Getting in the water.
How People are Using AI in 2025
  • Personal and Professional Support
  • Content Creation and Editing
  • Learning and Education
  • Technical Assistance and Troubleshooting
  • Creativity and Recreation
  • Research, Analysis, and Decision-making.
Look at your routines...
Where can AI help?
Repetitive tasks
Tasks that take a long time to do well
Tasks that benefit from a thought partner
AIRL Framework
Automate
Repetitive tasks
Save time for higher-value thinking
Improve
Edit, clarify, refine
From good to great…faster
Reimagine
Brainstorm and prototype
Think bigger, think different
Level-Up
Learn and reflect
Grow smarter with every interaction
Automate
Save time for higher-value thinking
Speech-to-Text
4X
Faster than typing
  • Works in any text field, desktop or mobile
  • Basic version is free
  • Doesn't have to be perfectly articulated
Meeting Intelligence
Automatic Transcription & Summaries

Centralizing your notes allows for easy search, summary, and analysis across conversations
Workflow Automation
Connect Your Apps
Automate repetitive tasks between tools

"Triggers" can be time or event based; "actions" might include sending an email, adding info to a spreadsheet, or more
Example

Pro Tip: Having trouble? Ask ChatGPT to walk you through it step-by-step and feed it screenshots to help it help you.
Model Context Protocols (MCPs)
Connect LLMs like ChatGPT to data and tools
Interrogate and Analyze Your Data
Ask questions across all your information
Take Action in External Tools
Use an LLM to create items in tools like Canva or Zapier
Work Smarter
Access context without manual searching

Available on paid plans of major AI platforms
Improve
From good to great—faster
Polishing or Adapting Text
Tools
  • ChatGPT
  • Claude
  • Gemini
  • Grammarly
  • Built into Microsoft Office, Google Docs, Notion
Tasks
  • Fixing grammar, spelling, flow
  • Refine for clarity and conciseness
  • Adapt for different audiences
  • Adjust tone and purpose
Example Prompts
"Fix grammar and improve flow."
"Streamline this memo, cutting it by 50% without losing key points or tone."
"Rewrite this policy section in plain language for faculty."
"Make this feedback more constructive and solution-focused."
"Check this document for hidden assumptions and bias and suggest alternatives."
"What critiques might a reader have of this document and how might I proactively address them?"
Tables
BEFORE: Unstructured Data
AFTER: Structured Data
Tables
BEFORE: Messy Data
AFTER: Standardized Data

Tip: describe what columns and rows you want OR create a mini exemplar for AI to copy
Presentations
Spend your time on substance, not style
Image Generators
Charts, Diagrams, Infographics
Visualize Complex Information
Transform data and concepts into compelling visuals
  • Generate charts from raw data
  • Create process diagrams automatically
  • Design infographics with AI assistance
  • Produce publication-ready graphics
Created with Claude.ai
Prototyping
My Sketch
Gemini's Mockup
Gemini's Nano Banana
"Generate a highly realistic superhero photo based on my selfie. I'm flying high above a city with natural photographic lighting, realistic skin texture, and detail."
Midjourney Animations
"Make me look like I'm flying into space."
"Vibe Coding" (No-code App Builders)
Build Apps by Describing Them
Lovable Demo
Reimagine
Think bigger, think different
Brainstorming
"Generate 10 creative or unconventional angles for a social media post about this topic."
"What are 5 very different ways I could structure this content?"
"Give me 3 radically different approaches to this problem."

Pro Tip: AI is a pattern matcher. Push it beyond the obvious by using terms like "creative," "unconventional," and "radically different."
Creative Perspectives
"You are an alien visiting Earth for the first time. Explain ______ as if you’re trying to make sense of this bizarre species."
"You’re a historian from the year 4025, piecing together what humans “probably” meant by ______ based on fragmented records."
"Provide the least obvious but still plausible interpretation of ______. No clichés allowed."
Consider Alternative Viewpoints
"How might my boss respond to this email?"
"How might school administrators perceive this decision?"
"How might a critic challenge this claim?"

Pro Tip: AI likes to agree. Give it explicit permission to disagree and be honest.
The Prompt
"A sweeping cinematic drone flyover of the Cliffs of Moher on Ireland's west coast. Early golden sunrise light hits the dramatic vertical cliffs, waves crash against the rocks far below, and seabirds glide through the frame. The camera glides smoothly along the cliff edge, capturing lush green grass rippling in the wind. Soft mist rises from the ocean, adding a magical, ethereal atmosphere. Hyper-realistic textures, natural lighting, and immersive ambient sound design."
Dublin's Ha'penny Bridge
"A cinematic shot of Dublin's Ha'penny Bridge at sunset. Warm orange light reflects off the River Liffey. People walk across the iconic white cast-iron bridge, some stopping to take photos. The camera tracks along the riverbank, capturing the movement of pedestrians, rippling water, and the soft glow of city lights turning on. Lively, atmospheric, modern Dublin vibe."
Real feedback…
"I LOVE the videos."
"Specifically the crafted ones? They feel very on topic and cohesive."
Level-Up
Grow smarter with every interaction
Is AI making us dumber?
Accelerating & Improving Learning
A lot of us see the potential…
Learning with "Study Modes"
AI as a learning companion accelerates skill development and deepens understanding
Why AI Works for Learning

YouTube does this really well. AI does this AND talks back.
Google's NotebookLM
Your AI Research Assistant
  • Upload documents and sources
  • Ask questions across materials
  • Generate summaries and insights
  • Create study guides, podcasts, videos, mind maps, flashcards, quizzes, infographics, and slide decks automatically
Reflection
Making Sense of Learning
Reflect with ChatGPT on what you just read
Processing Difficult Meetings
"What themes do you see here? What might I be missing?"
Weekly Review
"What patterns kept coming up? What should I prioritize next week?"

Why it matters: Processing matters. Sensemaking turns experience into wisdom.
AI for Equity
Democratizing Access
Personal Coach
Professional development guidance for everyone
Therapist
Mental health support available 24/7
Tutor
Personalized learning at your own pace
Financial Advisor
Expert money guidance regardless of net worth
Thank You
Questions? Let's continue the conversation.