Applying discipline expertise to AI Fluency

Shaping AI Fluency through Disciplinary Knowledge

Deep discipline knowledge is essential in shaping the AI fluency framework for any specific context. Interestingly, much of this development can happen without even touching an AI system. We are building human capacity and helping students—and ourselves—articulate what we know, how we know it, and why it matters. This foundational work is what makes us irreplaceable and keeps "humans in the loop."

You cannot partner with AI in your field effectively until you can articulate what quality looks like, how experts communicate, what problems matter, and what standards apply. To develop this crucial awareness, we can utilize the "4 Ds" framework: Discernment, Description, Delegation, and Diligence.

Discernment: Defining and Recognizing Quality

Discernment is one of the most important human competencies in the AI era. We need to deeply understand what "good" looks like in our field and possess a robust vocabulary to describe that quality to our AI assistants and to each other. While AI can generate endless content, only humans can judge what truly serves our specific goals and values.

Building Quality Criteria

Move beyond vague terms like "well-argued" or "creative." Work with colleagues and students to build detailed rubrics that capture deep quality markers.

• What specific moves make a philosophy paper compelling?
• What precise features make an engineering solution elegant? Document these criteria so students can internalize and apply them.

Analyzing Excellence

Collect outstanding work from your discipline, such as published papers, professional portfolios, and breakthrough solutions. Systematically analyze them with your students: mark up texts, diagram structures, and decode decisions. Create annotation guides that make expert thinking visible, helping students see what you see when you recognize excellence.

Diagnosing Failures

Equally important is studying what doesn't work. Collect flawed examples—failed arguments, buggy code, or ineffective designs—and examine them forensically.

• Where exactly does this proof break down?
• Why doesn't this solution scale?
• How does this fail to meet industry standards? Understanding failure modes builds the evaluative skills students need, whether working alone or with an AI agent.

Description: Making Tacit Knowledge Explicit

Every field has its own ways of communicating that embody its values and methods. Making this tacit knowledge explicit naturally results in strong description skills when prompting AI.

Mapping Disciplinary Products

Document the key products in your field with precision. Instead of just saying "lab reports," specify exact sections, conventions, and their underlying logic. Create templates that reveal why a methods section precedes results or what specific function the passive voice serves in scientific writing. Have students reverse-engineer professional outputs to understand their deep structure.

Revealing Expert Processes

Make the processes experts use to approach problems visible.

• How does a historian evaluate sources? Trace each micro-decision.
• How does a designer move from concept to prototype? Document each iteration. Have students interview practitioners and create flowcharts of expert thinking. This metacognitive work builds the awareness necessary for any collaboration, human or AI.

Naming Norms

Surface the behaviors that define your field. Scientists follow specific protocols for skepticism, replication, and peer review; artists experiment, critique, and revise. Define what it means to "think like a mathematician" in concrete, observable terms.

Delegation: Decomposing Work and Mapping AI Possibilities

Before students can work effectively with AI, they need to understand the components of the work itself. This allows them to decide when to delegate a task to an AI agent and when to retain human control.

The Anatomy of Problems

Teach students to break down challenges into component parts. Take a task like writing a literature review:

• Finding sources: Is this automatable?
• Evaluating relevance: How can AI augment this?
• Synthesizing arguments: Does this require human judgment?
• Crafting the narrative: Can this be done collaboratively?

Mapping AI Possibilities

Create detailed maps of what can be automated, augmented, or delegated.

• Automate: Routine tasks like data cleaning, initial drafts, or pattern recognition.
• Augment: Human-AI collaboration for design iteration, hypothesis generation, or literature synthesis.
• Delegate: AI agents acting independently to monitor experiments or filter information. Have students debate these boundaries using real-world examples.

Design Decision Trees

Create frameworks for when and how to involve AI. If analyzing qualitative data, where does software help versus hinder? If solving engineering problems, when does simulation replace physical prototyping? Build these trees through case studies, making delegation decisions explicit and defendable.

Diligence: Upholding Ethical Standards and Accountability

Every field has ethical standards and integrity requirements. Making these explicit prepares students for responsible practice in an AI-enhanced world.

Codifying Ethical Frameworks

Move beyond general academic integrity to discipline-specific ethics.

• If AI helps analyze patient data, how do we ensure privacy?
• If AI assists in legal research, how do we maintain client confidentiality? Build case studies and ethical decision matrices that students can apply to novel professional situations.

Clarifying Transparency

Work with students to document norms for disclosure. When must methods be fully disclosed? When is a process intentionally obscured? Create disclosure templates for common scenarios, such as citing AI assistance in a research paper versus a creative portfolio or a business report.

Co-creating Accountability

Draft classroom policies and honor codes together to address AI collaboration. Develop peer-review protocols that check for appropriate human oversight. This co-creation process builds buy-in and a deeper understanding of why accountability matters.

Conclusion: The Irreplaceable Human

This deep disciplinary work creates a powerful feedback loop. Students who can articulate quality standards can better evaluate any output, whether generated by a human or an AI. Those who understand disciplinary methods can guide any process more effectively.

Thinking through the lens of the 4Ds—Discernment, Description, Delegation, and Diligence—forces us to make tacit knowledge explicit. By making your expertise teachable, you are giving students the foundation they need to thrive. The future needs humans who can think critically, communicate clearly, collaborate wisely, and act responsibly. You are not preparing your students to be replaced by AI; you are preparing them to be irreplaceable.