Designing assignments for AI Fluency

Principles for Designing AI Fluency Assignments

When designing assignments that help students develop and demonstrate AI fluency, keep these three core principles in mind:

• Authenticity: Create assignments that mirror real-world AI collaboration rather than artificial exercises. Students should tackle genuine problems where AI partnership actually adds value.
• Iteration: Build opportunities for refinement that showcase growth over time. Single-shot assignments miss the iterative nature of fluent AI collaboration. Provide chances for students to refine, rethink, and try again.
• Pedagogical Transparency: Be clear that you are assessing the quality of the collaboration process and the students' own reflections, not just the final outputs. Students need to understand that you care about how they are learning to work with AI.

Outcome-Based Assignments

Outcome-based assignments focus on what students produce through AI collaboration. Effective approaches include:

1. Improving AI Outputs

Give students a mediocre AI output and have them transform it into something excellent. They must use discernment to identify problems and description to guide improvements. This raises awareness of discernment when using pre-existing content as a starting point.

2. Product Comparison

Have students use multiple AI systems or modes (such as extended thinking or image-to-image) to tackle the same task. They then analyze the different outputs, documenting which worked best for specific purposes and why. This builds platform awareness and teaches the core differences between various AI systems.

3. Constraint-Based Challenges

Give students specific output requirements regarding format, length, style, or audience. Students must work with AI to meet these constraints, submitting both their final product and the prompts used to achieve it. This develops description skills and can incorporate ethical, legal, or professional standards as additional constraints to build diligence.

4. Peer Product Review

Students set their own goals, create content with AI, and then swap with peers to critique the results against the stated goals. This builds product discernment and goal awareness while exposing students to different prompting approaches.

Process-Based Assignments

Process-based assignments aim to make the often private and fleeting human-AI interaction visible and assessable.

1. Annotated Chat Logs

Students submit records of their AI interactions, annotating key moments of insight, challenge, and growth. They should highlight turning points where their approach changed, breakthroughs in understanding, or how they recovered from a "hallucination" or failure.

2. Recorded Narrations

Students record themselves working with AI in real time using audio or screen capture. They narrate their decision-making process, explaining why they are making specific choices, what they are looking for in responses, and how they are evaluating the outputs.

3. Process Playbooks

Students build personal AI strategy guides. These reference documents record strategies and prompts for different tasks (e.g., research, creative writing, or problem-solving), noting which techniques work best in specific contexts.

4. AI-Assisted or Peer Debrief

Students share a chat log with an AI partner and ask the AI to discuss the interaction with them. Alternatively, this can be done with a human peer. Comparing insights from both AI and peer debriefs can provide greater depth in understanding their own delegation and description competencies.

Reflection-Based Assignments

Reflection is central to developing AI fluency and metacognitive awareness. To avoid "reflection fatigue," offer a diverse selection of approaches:

• Guided Inquiry: Provide specific questions for students to reflect on a particular assignment. Encourage honest critiques and complaints alongside positive experiences.
• Learning Journals: Students self-assess their 4D development (Delegation, Description, Discernment, and Diligence) at multiple points across the course. Encourage multiple modes of reflection, such as videos, infographics, or mind maps.
• Scenarios and Case Studies: Present realistic hypothetical scenarios (e.g., "Your boss wants to use AI for hiring decisions. What is your advice?") that challenge students to apply their learning to AI design, use, and impact.
• Personal Policy Statements: Have students create a personal policy statement based on their experiences. They develop their own values, strategies, and methods for fluent AI collaboration. This fosters empowerment and accountability.

Managing Content Volume

AI-related assignments can generate a significantly higher volume of content than traditional tasks. Here is how to manage the workload:

• Detailed, Deliverable-Based Rubrics: Use granular rubrics that allow you to quickly verify if students have completed specific deliverables well. This requires tight alignment between deliverables and learning outcomes.
• Emphasize Self and Peer Review: With proper guidance, students can observe their own development or that of their peers, condensing voluminous AI-generated outputs into manageable insights for faculty.
• Lightning-Round Conferences: Replace some written feedback with brief (5-minute) conferences where students present their key insights.
• Selective Sampling: You do not need to read every word of every chat log. Teach students to identify and flag key moments or "turning points" for your specific attention.

Summary

Designing AI fluency assignments is about creating opportunities for genuine learning through the practical application of the framework. Focus on authenticity, iteration, and pedagogical transparency.

• Outcome-based assignments assess description and discernment.
• Process-based assignments make human-AI interaction visible.
• Reflection-based assignments build metacognition through variety.
• Management strategies are essential to handle the increased volume of AI-generated content.