AI Literacy as a bottle neck

Decoding AI Literacy: Teaching Students to Think Critically About Their AI Use. This is not about whether or not we should be using AI but rather to think through how it's use must be a critical reflective process both for educators and for students.

Content

  • 1 Description of bottleneck
  • 2 Description of mental tasks needed to overcome the bottleneck
  • 3 Related scholarly work on this bottleneck
  • 4 People interested in this bottleneck
  • 5 Available resources
  • 6 References

1. Why AI Literacy is a Bottleneck

Students use AI tools in their coursework without critically examining their purpose, process, or implications of that use. They treat AI as an invisible tool rather than a choice requiring reflection. To frame it as an analogy, we are driving on a highway and we as teachers/educators/instructors see the AI exit but struggle to get to the exit, but with our students they don't even see the exit.

More Specific: Students default to using AI tools for assignments without considering: (1) whether AI is appropriate for the task, (2) how AI might be shaping their thinking, (3) what they might be missing by relying on AI, or (4) the ethical implications of their AI use

2. Description of mental tasks needed to overcome the bottleneck

Uncovering the Mental Move

To decode the expert mental moves, let us reword the bottleneck as a question: "How does one move from unconscious AI use to deliberate, critical engagement with AI tools?"

Expert Mental Moves When Using AI

Through the bottleneck writing tour method, we can identify what experts do differently
  1. Purpose Pause: Before using AI, experts stop to ask "What am I trying to accomplish and is AI the right tool?"
  2. Process Awareness: Experts consciously craft prompts, iterate, and document their interaction
  3. Output Evaluation: Experts don't accept AI output wholesale but evaluate, verify, and integrate it with their own knowledge
  4. Reflection Habit: Experts consider what they learned from the process versus what the AI provided
  5. Ethical Consideration: Experts think about citation, transparency, and the appropriateness of AI use for each context
Using AI to Decode This Bottleneck - AI can help us decode our own expert thinking about AI use
We can:
  • Use AI as an interviewing partner to probe your unconscious assumptions (maybe need a separate entry here or link to interview process in Decoding?)
    • Example prompt for self-decoding: "I'm trying to understand what I do automatically when I decide whether to use AI for a task. Can you interview me about a recent decision I made to use or not use AI, probing for assumptions I might not realize I'm making?"
  • Ask AI to challenge your thinking about when AI use is appropriate
  • Have AI generate scenarios that reveal hidden aspects of critical AI use

3. Modeling the Mental Move

The "AI Decision Tree" Metaphor

Back to our highway metaphor in the 'Why AI is a Bottleneck?' section: if we view AI use as a series of conscious forks in the road:

  • Each exit represents a choice about whether and how to use AI
  • Missing the exit means proceeding without thinking
  • Taking the exit requires slowing down and making deliberate decisions

Modeling Exercises:

1. Think-Aloud AI Use

Demonstrate your decision-making process for a real task:

  • "I need to write a paper draft introduction on ..."
  • "I could use AI to generate a first draft, but what would I miss?"
  • "Let me consider: What unique knowledge do I have that AI doesn't?"
  • "I'll use AI to check my logic flow, but write the content myself because..."

2. The "AI Interaction Journal"

Model keeping a record of:

  • Why you chose to use/not use AI
  • What prompts you tried and why you revised them
  • How you evaluated the output
  • What you kept, changed, or rejected and why

4. Practice and Feedback

Practice Sequence

How can we create scaffolded opportunities for students to practice critical AI engagement. This practice sequence can be used for educators/instructors in their own critical AI thinking

Week 1: Recognition Practice

  • Students identify AI use in their daily academic life
  • Document one unconscious AI use and reflect on it
  • Share in pairs: "I didn't realize I was using AI when..."

Week 2: Decision Point Practice

  • Give students identical tasks
  • Half use AI, half don't
  • Compare processes and outcomes, focusing on what each approach revealed

Week 3: Critical Prompt Development

  • Students develop prompts for the same task
  • Analyze how different prompts lead to different outputs
  • Reflect on what this reveals about AI as a tool

Week 4: Integration Practice

  • Complete a full assignment with documented AI decision points
  • Create an "AI use statement" explaining choices
  • Peer review focusing on critical thinking about AI use

AI literacy represents a fundamental conceptual threshold for students:

  1. Access and Leverage Research Methods: Without understanding AI fundamentals, students cannot effectively utilize increasingly AI-dependent research tools and methodologies.
  2. Critically Analyze Contemporary Social Phenomena: Many social interactions, inequalities, and power structures are now mediated through AI systems. Without literacy in this area, students lack the conceptual framework to properly analyze these phenomena. So, AI’s link to disinformation and being critical of media sources.

Key Bottleneck Areas

1. Methodological Understanding

Students who don't grasp basic AI literacy will struggle to:

  • Design research that accounts for algorithmic bias
  • Properly interpret AI-assisted data analysis
  • Evaluate the validity of AI-enhanced research methods

2. Theoretical Application

Limited AI literacy impedes students' ability to:

  • Apply existing social theories to algorithmic systems
  • Develop new theoretical frameworks that incorporate AI's social impacts
  • Connect traditional geographic (social science) concepts to emerging technological realities

3. Ethical Reasoning

Without AI literacy, students cannot:

  • Effectively assess ethical implications of AI deployment in social contexts
  • Develop frameworks for responsible AI use in social science research
  • Navigate complex issues of consent, privacy, and agency in algorithmic environments

Pedagogical Implications

  1. Threshold Concept Teaching: Design curriculum to explicitly address AI literacy as a threshold concept that transforms how students understand social phenomena
  2. Scaffolded Learning: Build progressive understanding from basic algorithmic concepts to complex socio-technical analysis
  3. Interdisciplinary Integration: Incorporate AI literacy across courses rather than isolating it to "tech" modules
  4. Practical Application: Provide hands-on experience with AI tools to overcome conceptual barriers through practice


Suggestions;

Give specific examples

Have a do and do not

No categories assignedEdit