Rethinking Learning Strategies in the Age of AI: Highlights from the 7th IIF Meeting

Adam’s presentation focused on doodling as a method of active learning that enhances visual literacy. He challenged the idea that doodling is a form of distraction, reframing it as a cognitive tool to process information and support memory retention.

Visual learning, he argued, is not new; it is deeply rooted in how humans make sense of the world through cave paintings, diagrams, and illustrations. Yet, in educational settings, we often undervalue drawing as a means of learning.

Theories such as dual-coding theory, cognitive load theory, and multimedia learning all support the premise that combining images with words improves understanding and retention. For instance:

• Visual elements can reduce cognitive overload when processing complex information.
• Doodling activates multiple modalities of learning: visual, kinesthetic, auditory, and reading/writing.
• It supports students in environments where content is dense, abstract, or fast-paced.

How to Doodle in Classrooms?

Doodling can be incorporated into teaching and learning in the following ways:

• As a lecturing aid: Live illustration to aid learning.
• As an engagement tool: Small group activities (e.g., drawing a concept with Post-Its).
• As a formative assessment:  Doodle-based exit tickets.

Adam introduced a structured yet flexible approach to incorporating doodling in teaching, as follows:

• Types of doodling: Texts with imagery, visual metaphors, structured layouts (e.g., flowcharts, mind maps).
• Example assignment: A 4-panel comic strip that features a dialogue between two characters, introduces and defines a concept from the lecture, and demonstrates understanding through visuals and student-generated language.
• Rubric ideas: Include creativity, conceptual accuracy, clarity of definition, and visual coherence.

Impact

Adam mentioned that students reported greater engagement and recall, particularly when visuals were tied directly to course concepts. Doodling was especially impactful in subjects requiring conceptual thinking or sequence-based content (e.g., biology, kinesiology). Challenges include the need for dual-display tech or document cameras and ensuring that students view doodling as a valid academic task, not a distraction.

Alejandro tackled one of the most pressing contemporary issues in writing-intensive courses: the misuse of AI. His approach reframes such a misuse as an opportunity for critical engagement and reflective learning rather than punishment.

Focusing on first-year writing courses, a space where AI use is increasingly prevalent, Alejandro showed how instructors can leverage AI-generated errors to promote deep thinking and classroom conversation. For instance, instead of alienating students or creating an adversarial classroom environment, Alejandro’s method seeks to:

• Strengthen students’ comprehension of course readings and writing skills.
• Encourage ethical reasoning around the use of AI.
• Foster a culture of inquiry rather than compliance.

He argued that traditional approaches often shut down dialogue, but his approach made AI part of the pedagogical process.

Turning Dishonesty into a Learning Opportunity

Alejandro provided strategies on how to implement his approach in classrooms:

• Paper feedback: When students submit papers with clear signs of AI use (e.g., factual inaccuracies, generic summaries, false citations), Alejandro flags them, not to accuse, but to generate prompts for class discussion.
• Classroom activities: Turn the mistakes into in-class writing and discussion prompts and invite students to rethink them.
• Individual reflection: Students are given time to reflect on the in-class discussion, asking questions, such as:
  • Why were you able to identify AI’s mistake?
  • What do you understand that the AI does not?

Impact

Alejandro noticed a significant difference in the suspected use of AI. In Unit 1, 7 students submitted AI-generated content, but by Unit 2, only one did. More importantly, students became less defensive and more open to discussing AI use, academic integrity, and what it means to truly understand a text. Alejandro noted that while the approach thrived in his context and discipline (obscure texts, lesser-known writers), it has broader potential, especially when instructors frame AI as a tool for thinking rather than a threat.

Shubha presented a well-structured strategy for teaching numerical methods in physics. The strategy focused on active learning and formative assessment rather than high-stakes exams. The course aims to help students navigate the “third pillar” of physics: simulation and computation, alongside theory and experimentation.

Challenges in computational physics classes are many, such as:

• Students enter with varying coding backgrounds.
• Learning paces differ widely.
• Easy access to AI-generated code can mask gaps in understanding.

Shubha emphasised that understanding the problem is at the heart of problem-solving, and no amount of correct output can substitute for conceptual clarity.

What is the Strategy?

Shubha structured the course in such a way that it supports different learning preferences and promotes responsibility:

• Twice-weekly sessions: Each session begins with a short lecture (45–60 min), followed by hands-on lab time.
• Weekly exercises (formative):
  • Goal: Students to reach 100 points/week through a mix of easy/hard problems.
  • Detailed handouts and example code are provided to support students in choosing their approach.
  • TAs and instructors conduct “check-ins” to assess not just completion but also comprehension.
• Summative projects: Four major homework projects plus a final presentation.

To promote student learning responsibility, Shubha made academic integrity clear and addressed it early on. For example, students were informed of the academic honesty statement and emphasised the importance of understanding over copying and pasting, especially when it comes to coding. They were also instructed that proper attribution is required for any outside or AI-generated code.

Impact

Shubha observed that students became more comfortable working independently and more thoughtful about how they use AI tools. Subha’s model shows that with structure and active engagement, even technically demanding courses can foster ethical and self-directed learning.