What it looks like: In level one, students are expected to complete the work without any AI assistance. These assessments are designed to protect the specific kind of learning that benefits from students working through something on their own or with their peers. Assignment or syllabus directions should note any exceptions for built-in tools. For example, the AI-powered spell-check and grammar suggestions found in Office 365 tools are enabled by default. We recommend explicitly listing these as exceptions, or if your course needs are different, asking students to acknowledge their use.

When it makes sense:

• The assignment is designed to capture an authentic, unmediated demonstration of what students individually know, think, or can do.
• Students are still developing foundational skills that AI might shortcut.
• Professional or accreditation standards require demonstrated independent performance.

Assignment Examples

In-Class Diagnostic Reflection    

Quick, timed, low-pressure assignments can give you immediate insight into students’ genuine understanding of or takeaways on a given topic.

• At the start of class, spend 8–10 minutes sketching a concept map of everything you remember about [course-specific topic]. Work from memory, without referencing your notes or looking anything up. Include key terms, relationships, and any connections to other concepts you’ve encountered. Don't worry about getting it perfect; the goal is to capture what you already know and where your thinking currently stands.
• Before responding to today’s iClicker question, take 2–3 minutes to discuss your reasoning with your partner or group. Focus on why you chose your answer, not just what it is. If you’re unsure, talk through your uncertainty together.

Sketchbook Assignments   

An assignment requiring iterative, physical drafting can often help students develop their ideas and demonstrate the progression of their learning. Key features of sketchbook assignments include iteration and annotation to make the thinking transparent and combine the visual affordances of an open space with written language (images with captions, mind and concept maps, storyboards, a simple idea redrawn from different perspectives).

• Over the course of the semester, we’ll be reading several religious texts from different faith traditions on the concept of justice. Each week, please create a visual or mind map of how the week’s text defines justice. Your mind map or visual should make clear how the text defines justice, what factors impact justice, who is included or excluded in the text’s definitions, and what would it look like for that concept of justice to be applied in your life today? Be sure to annotate your visuals to connect evidence from the reading and make your reasoning clear. In the final entry of your sketchbook, create a combined map that demonstrates the relationships and incongruencies across the different faith traditions we’ve discussed.
• Create a design notebook that proposes a restoration plan for a degraded riparian region that balances ecological health, cultural practices from local communities, and long-term sustainability. Be sure to record your research and observations, questions, and ideas as you iterate on your solution throughout the semester. Someone examining your sketchbook should be able to follow your logic and understand how you arrived at your proposed solution.

Oral Presentation, Dialogue, or Defense   

Oral presentations are inherently difficult to outsource to AI; they also develop the type of on-the-spot thinking and communication skills that professional contexts can demand.

• You’ll have 10 minutes to present the core argument of your final paper, followed by 10 minutes of questions. Be prepared to explain your reasoning, justify your sources, and respond to counterarguments.

What it looks like: Students may use AI tools during the planning, brainstorming, or outlining phase, but the actual work, such as the writing, analysis, or argument, should be their own. At this level, AI is positioned as a thinking partner, not a substitute for students’ own voice and reasoning.

When it makes sense:

• You want students to develop their own ideas but provide permission to use AI as a sounding board.
• Students are novices who might benefit from AI-generated prompts or outlines to model possible ways to organize their thinking; however, they still need to produce the substantive work themselves.
• You want to teach students to recognize where AI-generated ideas are generic or incomplete.

Assignment Examples

Podcast with AI Task Creator   

AI can be a useful planning partner, but its real value lies in what it prompts students to notice: the gaps, assumptions, and oversights that only a critical eye can catch and correct. For novice users, it might be appropriate to build that critical awareness collectively.

• Prompt an AI tool to create a step-by-step task list for creating a short podcast episode on [course-specific topic]. Ask it to break each task into steps that can be completed in 20–30 minutes. Then, review the list, considering what you would add, remove, or change. Submit the AI-generated list and a few sentences reflecting on what you noticed. We’ll crowdsource the examples in class to help you refine a project plan for your final assignment.

Research Question Development   

In the following task, AI works to surface assumptions, complications, or framings that students may not have considered. Students remain in charge, however, deciding what to keep, what to discard, and why that distinction matters for their research project.

• Before our next class, use an AI tool to help you refine your research question. Start with your initial question, then ask the AI to identify potential gaps, complications, or alternative framings. Submit your original question, as well as your revised question. Include a short explanation of what shifted in your thinking and why the revised research question is a stronger option for your project.

What it looks like: Students engage with AI throughout the task, but the expectation is that student thinking should drive every decision.  Students are expected to critically evaluate what AI produces, push back where it falls short, and make intentional choices about what to keep and what to discard.

When it makes sense:

• Students have enough background knowledge to recognize when AI outputs are wrong, generic, or missing important nuance
• The learning goal involves developing critical judgment and meta-awareness about AI limitations
• You want students to practice the kind of AI-assisted workflow they’ll encounter in professional settings.

Assignment Examples

Place-based Comparison:  

Local, place-based knowledge underscores AI’s tendency toward abstraction. Noticing when AI-generated descriptions don’t match or adequately convey the nuances of lived experiences builds essential critical skills.

• Ask an AI tool to describe what [a social issue relevant to your course] looks like in [local community or regional context]. Then write a 200-word response that identifies at least one specific way the AI’s description doesn't match what you know about your community or what you’ve learned in this course. Use one concept from this week’s reading to frame your response.

Side-by-side Analysis   

This type of assignment fosters the development of analytical skills and meta-cognitive awareness. By positioning AI as an intellectual foil, students are challenged to question and refine their own ideas, as well as AI outputs.

• Write your own analysis of [case study / text / dataset]. Next, ask an AI tool to analyze the same material. Place the two analyses side by side. In a 300–400 word reflection, identify: 1) where the AI’s analysis aligns with yours; 2) where it diverges and why; and 3) what the AI overlooked, based on your understanding of course content and disciplinary knowledge.

What it looks like: Students use AI extensively and openly, but the emphasis is on strategic, purposeful direction of AI output. At this level, the student assumes the role of expert by evaluating, curating, and steering AI toward a specific, predetermined goal. Importantly, the student’s own knowledge sets the limits here. Students are judging and refining AI output against what they already know, which means the quality of their direction is bounded by their existing understanding.

When it makes sense:

• Students have strong enough disciplinary knowledge to catch errors, bias, or limitations in AI outputs.
• The learning goal involves cultivating a culture of openness surrounding AI use, supported by the intentional development of AI literacy skills.
• The task mirrors professional contexts where AI-assisted work is becoming a norm.

Assignment Examples

Comparative Analysis of AI Platforms   

By comparing AI platforms, students’ awareness of the differences across tools deepens.

• Present the same professional scenario to three different AI tools (for example, ChatGPT, Claude, and Gemini). Compare their recommendations. Then, drawing on course concepts and professional standards, write a 500-word analysis of which response you’d trust most, which you’d trust least, and what that comparison reveals about the limitations and affordances of these tools in this professional field.

AI-Generated Report with Critique   

Although generative AI is becoming a common feature of professional practice in fields such as communications and public health, the ability to critically assess AI outputs through a disciplinary lens remains an essential professional competency.

• Use an AI tool to produce a media package for client X. Write a 500-word critical analysis of the AI-generated content that addresses the following: the accuracy of its content, noting any factual errors or unsupported claims; the appropriateness of its tone and structure for both the client and their target audiences; and any identifiable biases or limitations in how the media package was presented. In addition, document your working process, revisions made, and your reasoning behind key decisions.

What it looks like: Students and instructors co-investigate what AI can and cannot do within a specific discipline or context. Unlike Level 4, the goal is not to direct AI toward a known outcome, but to explore open questions, including novel applications where neither the student nor the instructor knows in advance what AI will produce or reveal. At this level, AI functions less as a tool to be managed and more as a collaborator in an unscripted inquiry. Work at this level can potentially contribute to broader conversations about AI in professional practice.

When it makes sense:

• Students are advanced enough to engage with the boundaries and emerging questions within the disciplinary field.
• The assignment is framed as an investigation, not a demonstration of existing competency.
• The stakes are structured so that risk-taking is rewarded, not penalized.

Assignment Examples

Field-Specific AI Audit   

In this task, students generate new knowledge about AI’s affordances and limitations within their specific domain.

• Choose a task central to legal practice and use an AI tool to perform that task across five different scenarios involving case summarization and the identification of key legal arguments. Rather than evaluating a known outcome, treat the AI as a collaborator in an open inquiry: Observe what it surfaces, misses, or constructs in ways you did not anticipate. Pay particular attention to moments where the AI hallucinates citations, perpetuates bias, produces poor quality legal summaries, or generates arguments that appear authoritative but may not withstand scrutiny. Use those moments as entry points, experimenting with how refined prompting, reframing, or iterative feedback might shift or improve its outputs, and what that process itself reveals. Your conclusion should address what this collaboration, including its failures and recoveries, reveals about the nature of legal reasoning, the evolving role of the lawyer in the age of AI, and how working this way might contribute to broader conversations about AI in legal practice.

Boundary-Testing Inquiry   

Students explore AI limits in the following task, positioning themselves as contributors to the field's emerging conversation about AI.

• Explore what it would mean for an expert clinical practitioner, positioned at the intersection of medicine and AI, to identify, interrogate, and begin correcting bias in a clinical decision-support tool suspected of producing differential recommendations across patient populations. Use an AI tool as part of this inquiry, reflecting on what it surfaces, what it cannot see, and what that reveals about the limits of algorithmic self-awareness. Although this is not a question with a settled answer, your conclusion should address what genuine clinical accountability for biased AI looks like in practice, including who bears responsibility, at what stage intervention becomes possible, and what this process reveals about the gap between the promise of objective medical AI and the reality of its deployment.

This week, we will be practicing how to evaluate AI outputs. Go to an Generative AI tool (you can access Co-Pilot with your UNM credentials) and ask it a complex, multi-layered question about the acequia systems of New Mexico. For example, avoid asking a general question like: "What is an acequia?" Instead, ask a specific question that extends this week's reading and/or connects to your lived experience, such as "How do traditional acequia governance structures and the role of the mayordomo come into conflict with contemporary state water rights during severe droughts in Northern New Mexico?" Step 1: Copy and paste the AI's exact response into your discussion post. Step 2: Create a short (<5 minutes) video where you question and analyze the AI response based on your own thinking and this week's reading as evidence. What's AI getting right, and what contexts are it missing? Step 3: In your replies to two peers, start to synthesize AI's contributions and setbacks to this week's conversation. Where has it added to the conversation, and where would it have steered us wrong if not for further questioning?

This week, we are examining the modern challenges facing acequia communities. You have been assigned a specific stakeholder role for this discussion:

• a traditional mayordomo (acequia manager)
• a newly arrived commercial farmer
• a NM state water rights attorney.

Use AI to brainstorm 2–3 arguments, legal concerns, and historical priorities for your role. Produce one original post of 150 words applying AI ideas to the case study. Synthesize case details in your own voice; paraphrase sources and cite the study. In your replies, you must respond to the position statements of a peer representing a different stakeholder and attempt to arrive at a compromise.

For this week's discussion, generative AI is strictly prohibited. To connect our historical readings to the present day, go into your own community and take an original photograph of local water infrastructure (such as an active acequia, a modern irrigation ditch, a neighborhood storm drain, or a dried riverbed). Post your photograph to the discussion board and use Kaltura to record a 90-second audio or video reflection. In your recording, explain how your photograph either reflects or contrasts with the traditional communal water practices we read about this week. Reply to two peers via video or text. Engage directly with the original reflection, referencing a specific element of the peer's photo, narration, or interpretation that connects to an idea or theme in this week's reading.