WAIRED
LEARN WITH(OUT) AI
03-04-2025 BY SIMONE DE GIJT
Slidedeck:
GLOSSARY
• Jagged Frontier
• More terms:
• Check your favorite AI assistant
• https://www.coursera.org/resources/ai-terms
JAGGED FRONTIER
Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality
- Harvard Business School
Refers to the unpredictable and
ever-changing limits of what AI
can and cannot do
#wAIred
WHAT CAN AI DO FOR US?
CLOSING THE GAP
22.12%
4.46%
Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality
- Harvard Business School
#wAIred
COPYING ANSWERS
A.L. Glass & M. Kang (2020): Fewer students are benefiting from doing their homework: an eleven-year study, Educational Psychology,
DOI: 10.1080/01443410.2020.1802645
#wAIred
#wAIred
AI TOOLS
Type of AI What can it do? Brands
AI Assistants Chat
Plan
Summarize
..
ChatGPT
Copilot
Claude
..
Type of AI What can it do? Brands
AI Assistants Chat
Plan
Summarize
..
ChatGPT
Copilot
Claude
..
Code generation
embedded in IDEA
Code completion
Explain
Simplify
..
Jetbrains AI Assistant
Github Copilot Tabnine
..
Type of AI What can it do? Brands
AI Assistants Chat
Plan
Summarize
..
ChatGPT
Copilot
Claude
..
Code generation
embedded in IDEA
Code completion
Explain
Simplify
..
Jetbrains AI Assistant
Github Copilot Tabnine
..
Workspaces Query
Vocalize
..
NotebookLM
Notion
..
Type of AI What can it do? Brands
AI Assistants Chat
Plan
Summarize
..
ChatGPT
Copilot
Claude
..
Code generation
embedded in IDEA
Code completion
Explain
Simplify
..
Jetbrains AI Assistant
Github Copilot Tabnine
..
Workspaces Query
Vocalize
..
NotebookLM
Notion
..
Supporting tools Visualization
..
Napkin.ai
..
#wAIred
NAPKIN.AI
THE IMPACT OF AI ON
LEARNING CAPABILITIES
A RISK ASSESSMENT
• Overreliance:
• Users may not engage deeply with the material
• Superficial understanding of programming concepts
• Novice learners may not develop their own problem-solving skills
• Bias and Errors:
• Risk of low code quality as AI is always choosing the most predictable solution
• Users’ judgement may get clouded by the preferences and biases of the AI
• Falsehoods may be saved to the long-term memory and become the base for future decisions and knowledge
• Illusion of understanding:
• Accept answers without (the skills to) evaluating the accuracy
• Novice learners may not develop their own critical thinking skills
• Falling asleep at the wheel
#wAIred
FALLING ASLEEP AT THE WHEEL
Let AI take over instead of using
it as a tool. Which hurts human
learning, skill development and
productivity.
#wAIred
Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality
- Harvard Business School
#wAIred
OpenAI (2023): GPT-4 Technical Report, DOI: 10.48550/arXiv.2303.08774
IS AI REALLY THAT FLAKY?
#wAIred
TESTING BECOMES MORE IMPORTANT
• Testing encourages you to think critically about the code and
understand its functionality.
• Ensures that the code functions as intended and meets the required
specifications.
• Helps in catching errors and bugs that AI might introduce or overlook.
#wAIred
AN OPPORTUNITY ASSESSMENT
• Idea Generator on Steroids:
• AI can generate a wide range of ideas and solutions quickly, sparking creativity
• Offering different perspectives so users can improve their critical thinking skills
• More opportunities for users to reflect on their progress and choices
• Personalized Learning:
• Offering instant help with problems, explanations or examples
• AI can adapt to individual learning styles & needs
• Can increase student participation and motivation
• Reducing cognitive load
• Freeing up cognitive resources for higher-order thinking
• Allowing users to focus on more complex and critical aspects of their learning
#wAIred
BALANCING AI AND
HUMAN INTELLIGENCE
STEP 1. FIND YOUR SOURCE OF TRUTH
We need subject matter expertise to:
• Think critically
• Problem solve
• Understand abstract concepts
• Reason through novel problems
• Evaluate the AIs output
#wAIred
STEP 2. INVITE AI TO EVERYTHING YOU DO
• Learn the shape of the jagged frontier
• Learn how to refine prompts
• Imagine your AI collaborator as an
infinitely fast intern, eager to please
but prone to bending the truth.
#wAIred
STEP 3. GO CENTAUR MODE
• Clear line between person and
machine
• Assign clear, stand-alone tasks
to the AI which are verifiable
#wAIred
LEARNING EPISODE
1.Receive learning material
2.Ask AI questions about the learning for deeper understanding
3.Practise session
4.AI grades the performance
5.AI advices how to continue
#wAIred
STEP 4. GO CYBORG MODE
• Blend machine and person
• AI doesn’t just do delegated tasks
• The user and the AI move back
and forth over the jagged frontier
#wAIred
EFFECT OF AI
CODE GENERATORS
ON SUPPORTING
NOVICE LEARNERS
• Code-Authoring Performance:
Codex users had a 1.15x increased completion
rate and 1.8x higher scores.
• Manual Code-Modification:
No decrease in performance for Codex users.
• Retention:
Slightly better performance in post-tests for
Codex users, though not statistically significant.
Kazemitabaar, Chow, Ka To Ma et all. (2023): Studying the effect of AI Code Generators on Supporting Novice Learners in Introductory Programming, CHI,
DOI: https://doi.org/10.1145/3544548.3580919
#wAIred
KEY TAKE AWAYS
AI can generate creative ideas quickly, enhancing decision-making and problem-solving.
Overreliance on AI can lead to superficial understanding and hinder critical thinking skills.
AI tools can personalize learning experiences, adapting to individual needs and styles.
Biases and errors in AI outputs can affect judgment and decision-making.
Balancing AI with human intelligence is crucial for effective learning and productivity.
#wAIred
SOURCES
- Dell’Acqua, McFowlan III, Mollick et all. (2023): Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge
Worker Productivity and Quality. Harvard Business School.
- Glass & Kang (2020): Fewer students are benefiting from doing their homework: an eleven-year study, Educational Psychology, DOI:
10.1080/01443410.2020.1802645
- Kazemitabaar, Chow, Ka To Ma et all. (2023): Studying the effect of AI Code Generators on Supporting Novice Learners in Introductory Programming, CHI,
DOI: https://doi.org/10.1145/3544548.3580919
- Lin & Chen (2024): Artificial intelligence (AI) -integrated educational `zapplications and college students’ creativity and academic emotions: students and
teachers’ perceptions and attitudes, BMC Psychology, DOI: https://doi.org/10.1186/s40359-024-01979-0
- Mollick (2024). Co-Intelligence: Living and Working with AI.
- OpenAI (2023): GPT-4 Technical Report, DOI: 10.48550/arXiv.2303.08774
- Spector & Ma (2019). Inquiry and critical thinking skills for the next g`zeneration: from artificial intelligence back to human intelligence. Smart Learning
Environments, DOI: https://doi.org/10.1186/s40561-019-0088-z
- Yi Wu (2024): Critical Thinking Pedagogics Design in an Era of ChatGPT and Other AI Tools — Shifting From Teaching “What” to Teaching “Why” and “How”,
DOI: https://doi.org/10.20849/jed.v8i1.1404
- Zviel-Girshin (2024): The Good and Bad of AI Tools in Novice Programming Education. Education sciences, DOI: https://doi.org/10.3390/educsci14101089
#wAIred
LinkedIn:
Simone de Gijt
Slidedeck:
#wAIred
Thank you for your attention!

Internet Download Manager (IDM) 6.42.40 Crack Download

  • 1.
    WAIRED LEARN WITH(OUT) AI 03-04-2025BY SIMONE DE GIJT Slidedeck:
  • 2.
    GLOSSARY • Jagged Frontier •More terms: • Check your favorite AI assistant • https://www.coursera.org/resources/ai-terms
  • 3.
    JAGGED FRONTIER Navigating theJagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality - Harvard Business School Refers to the unpredictable and ever-changing limits of what AI can and cannot do #wAIred
  • 4.
    WHAT CAN AIDO FOR US?
  • 5.
    CLOSING THE GAP 22.12% 4.46% Navigatingthe Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality - Harvard Business School #wAIred
  • 6.
    COPYING ANSWERS A.L. Glass& M. Kang (2020): Fewer students are benefiting from doing their homework: an eleven-year study, Educational Psychology, DOI: 10.1080/01443410.2020.1802645 #wAIred
  • 7.
  • 8.
  • 9.
    Type of AIWhat can it do? Brands AI Assistants Chat Plan Summarize .. ChatGPT Copilot Claude ..
  • 10.
    Type of AIWhat can it do? Brands AI Assistants Chat Plan Summarize .. ChatGPT Copilot Claude .. Code generation embedded in IDEA Code completion Explain Simplify .. Jetbrains AI Assistant Github Copilot Tabnine ..
  • 11.
    Type of AIWhat can it do? Brands AI Assistants Chat Plan Summarize .. ChatGPT Copilot Claude .. Code generation embedded in IDEA Code completion Explain Simplify .. Jetbrains AI Assistant Github Copilot Tabnine .. Workspaces Query Vocalize .. NotebookLM Notion ..
  • 12.
    Type of AIWhat can it do? Brands AI Assistants Chat Plan Summarize .. ChatGPT Copilot Claude .. Code generation embedded in IDEA Code completion Explain Simplify .. Jetbrains AI Assistant Github Copilot Tabnine .. Workspaces Query Vocalize .. NotebookLM Notion .. Supporting tools Visualization .. Napkin.ai ..
  • 13.
  • 14.
    THE IMPACT OFAI ON LEARNING CAPABILITIES
  • 15.
    A RISK ASSESSMENT •Overreliance: • Users may not engage deeply with the material • Superficial understanding of programming concepts • Novice learners may not develop their own problem-solving skills • Bias and Errors: • Risk of low code quality as AI is always choosing the most predictable solution • Users’ judgement may get clouded by the preferences and biases of the AI • Falsehoods may be saved to the long-term memory and become the base for future decisions and knowledge • Illusion of understanding: • Accept answers without (the skills to) evaluating the accuracy • Novice learners may not develop their own critical thinking skills • Falling asleep at the wheel #wAIred
  • 16.
    FALLING ASLEEP ATTHE WHEEL Let AI take over instead of using it as a tool. Which hurts human learning, skill development and productivity. #wAIred
  • 17.
    Navigating the JaggedTechnological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality - Harvard Business School #wAIred
  • 18.
    OpenAI (2023): GPT-4Technical Report, DOI: 10.48550/arXiv.2303.08774 IS AI REALLY THAT FLAKY? #wAIred
  • 19.
    TESTING BECOMES MOREIMPORTANT • Testing encourages you to think critically about the code and understand its functionality. • Ensures that the code functions as intended and meets the required specifications. • Helps in catching errors and bugs that AI might introduce or overlook. #wAIred
  • 20.
    AN OPPORTUNITY ASSESSMENT •Idea Generator on Steroids: • AI can generate a wide range of ideas and solutions quickly, sparking creativity • Offering different perspectives so users can improve their critical thinking skills • More opportunities for users to reflect on their progress and choices • Personalized Learning: • Offering instant help with problems, explanations or examples • AI can adapt to individual learning styles & needs • Can increase student participation and motivation • Reducing cognitive load • Freeing up cognitive resources for higher-order thinking • Allowing users to focus on more complex and critical aspects of their learning #wAIred
  • 21.
  • 22.
    STEP 1. FINDYOUR SOURCE OF TRUTH We need subject matter expertise to: • Think critically • Problem solve • Understand abstract concepts • Reason through novel problems • Evaluate the AIs output #wAIred
  • 23.
    STEP 2. INVITEAI TO EVERYTHING YOU DO • Learn the shape of the jagged frontier • Learn how to refine prompts • Imagine your AI collaborator as an infinitely fast intern, eager to please but prone to bending the truth. #wAIred
  • 24.
    STEP 3. GOCENTAUR MODE • Clear line between person and machine • Assign clear, stand-alone tasks to the AI which are verifiable #wAIred
  • 25.
    LEARNING EPISODE 1.Receive learningmaterial 2.Ask AI questions about the learning for deeper understanding 3.Practise session 4.AI grades the performance 5.AI advices how to continue #wAIred
  • 26.
    STEP 4. GOCYBORG MODE • Blend machine and person • AI doesn’t just do delegated tasks • The user and the AI move back and forth over the jagged frontier #wAIred
  • 27.
    EFFECT OF AI CODEGENERATORS ON SUPPORTING NOVICE LEARNERS • Code-Authoring Performance: Codex users had a 1.15x increased completion rate and 1.8x higher scores. • Manual Code-Modification: No decrease in performance for Codex users. • Retention: Slightly better performance in post-tests for Codex users, though not statistically significant. Kazemitabaar, Chow, Ka To Ma et all. (2023): Studying the effect of AI Code Generators on Supporting Novice Learners in Introductory Programming, CHI, DOI: https://doi.org/10.1145/3544548.3580919 #wAIred
  • 28.
    KEY TAKE AWAYS AIcan generate creative ideas quickly, enhancing decision-making and problem-solving. Overreliance on AI can lead to superficial understanding and hinder critical thinking skills. AI tools can personalize learning experiences, adapting to individual needs and styles. Biases and errors in AI outputs can affect judgment and decision-making. Balancing AI with human intelligence is crucial for effective learning and productivity. #wAIred
  • 29.
    SOURCES - Dell’Acqua, McFowlanIII, Mollick et all. (2023): Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality. Harvard Business School. - Glass & Kang (2020): Fewer students are benefiting from doing their homework: an eleven-year study, Educational Psychology, DOI: 10.1080/01443410.2020.1802645 - Kazemitabaar, Chow, Ka To Ma et all. (2023): Studying the effect of AI Code Generators on Supporting Novice Learners in Introductory Programming, CHI, DOI: https://doi.org/10.1145/3544548.3580919 - Lin & Chen (2024): Artificial intelligence (AI) -integrated educational `zapplications and college students’ creativity and academic emotions: students and teachers’ perceptions and attitudes, BMC Psychology, DOI: https://doi.org/10.1186/s40359-024-01979-0 - Mollick (2024). Co-Intelligence: Living and Working with AI. - OpenAI (2023): GPT-4 Technical Report, DOI: 10.48550/arXiv.2303.08774 - Spector & Ma (2019). Inquiry and critical thinking skills for the next g`zeneration: from artificial intelligence back to human intelligence. Smart Learning Environments, DOI: https://doi.org/10.1186/s40561-019-0088-z - Yi Wu (2024): Critical Thinking Pedagogics Design in an Era of ChatGPT and Other AI Tools — Shifting From Teaching “What” to Teaching “Why” and “How”, DOI: https://doi.org/10.20849/jed.v8i1.1404 - Zviel-Girshin (2024): The Good and Bad of AI Tools in Novice Programming Education. Education sciences, DOI: https://doi.org/10.3390/educsci14101089 #wAIred
  • 30.