The Description-Discernment loop

Introduction to the Description Discernment Loop

Hello, I'm Joe Feller. Let's explore the Description Discernment Loop, the part of the AI fluency framework that students often find the most immediately applicable and useful. While the Delegation Diligence Loop handles the bigger picture, this loop is about the moment-to-moment craft of building collaborative environments where humans and AI can work together effectively.

When students think about AI skills or literacy, they tend to think about clever hacks and prompting tricks. This offers a more sophisticated and enduring approach, preparing students for professional AI interaction now and into the future.

Beyond Prompt Engineering: Building Cognitive Environments

A lot of AI interaction advice focuses on crafting the perfect prompt. While prompt engineering is useful, thinking in terms of the Description Discernment Loop invites us to think bigger—about conversations rather than just commands. This fundamentally changes the mindset of how students approach AI.

We're teaching them not to simply engage in an input-output exchange, as if they were typing commands into an operating system or app. Instead, we're teaching them to build context and a shared understanding with AI about the world and the tasks at hand. On a practical level, this change of mindset allows students to communicate more effectively with AI both now and into the future, without copy-pasting prompts from others or following very specific step-by-step guides.

Furthermore, we're teaching them to build the overarching cognitive environment in which they interact with AI. The context we give an AI system directly impacts how it processes its training data and how it responds to our prompts. Similarly, when you have a great conversation with a colleague, you build on each other's ideas, develop shorthand, and know when to push back or encourage. You learn to think together. That is what we mean by a cognitive environment.

These environments may include:

• Shared vocabulary and references that evolve over time.
• Well-defined goals, values, processes, and methods.
• Established patterns of interaction that enable both human and AI to perform to the best of their ability.
• Mechanisms for building on previous interactions.

This is real collaboration where knowledge deepens, processes evolve, and innovative thinking emerges. You definitely can't get there with a single prompt, no matter how well-engineered.

The Three Elements: Product, Process, and Performance

The Description Discernment Loop includes three interconnected elements: Product, Process, and Performance. These operate like different lenses for understanding the same collaborative process.

Product-Focused Collaboration

Product focus starts with clearly articulating what we're trying to create together and then rigorously evaluating the result. This isn't a one-time specification; it's an evolving understanding of what "good" looks like. In many cases, students will be improving their own understanding alongside the AI's. This is when we move beyond automation and into augmentation.

For example, a student might begin with a broad prompt: "Help me write about social media and teenagers." This generic prompt yields a generic response. Through discernment, the student realizes this isn't what they wanted. Their next description becomes more specific: "I'm exploring how teens present different versions of themselves across platforms." Each version improves the final item and provides greater mental clarity to the student. Eventually, they might arrive at: "Craft this analysis through the lens of how teens present different versions of themselves across platforms. We want readers to understand the psychological toll of maintaining these different personas."

Process-Focused Collaboration

Process-focused collaboration is about developing shared approaches to thinking and problem-solving. Consider a student asked to analyze documents. They might start with something vague like "Analyze this step-by-step." After trial and error, they develop richer instructions:

"When we analyze each document, let's start by identifying the author's context and discussing how this might influence the content. After that, we'll look at the content itself—both what is said and any notable omissions. Then we will consider multiple interpretations before settling on our conclusions."

Process discernment evaluates whether the collaboration is working by asking: Is the process uncovering insights we might miss alone? Is it becoming more sophisticated over time?

Performance-Focused Collaboration

Performance is about the relationship between the student and the AI assistant. We want to create dynamics that enable both to contribute their best thinking. A student might initially approach AI formally, but iterations lead to something more personal:

• "Act as a kind but rigorous skeptic. When I propose ideas, help me find the weak points in a way that builds my confidence."
• "Use analogies from music to help make these concepts stick."
• "This is a creative task, so let's keep some humor in the work. I think more creatively when I'm relaxed."

Strategies for Developing AI Collaboration Skills

How do we help students develop these skills?

1. Go beyond the single prompt: Design assignments that require multiple interactions over time in a single chat, where context accumulates. Then, design assignments across several chats so students understand what context is preserved and what needs to be re-established.
2. Share your own process: Let students see your AI collaborations. Show how your understanding deepens, how you build shared language, and how you adjust interaction dynamics.
3. Encourage reflection: Have students document the evolution of their collaboration, not just the outputs. How did the "3 Ps" (Product, Process, and Performance) interact in practice?

Recognizing Signs of Strong Collaboration

We need to help students recognize two positive signs of strong collaboration:

• Shared Language: As context deepens, shorthand references to complex ideas and reusable solution patterns from earlier conversations become common. This makes collaboration easier and richer.
• An Exploration Mindset: Initial interactions might feel like programming, but as trust builds, the style evolves from a rigid "do X and then do Y" into something more flexible and context-aware. Students can go beyond getting what they asked for to discovering what they didn't know to ask.

Conclusion: The Integrated Framework

The Description Discernment Loop teaches students that meaningful AI collaboration means building environments where human intelligence and AI combine in powerful ways.

• Humans contribute context-rich judgment, creativity, ethical reasoning, and emotional intelligence.
• AI partners offer vast knowledge synthesis, pattern recognition, and the ability to hold complexity without being overwhelmed.

The two loops work in concert: The Delegation Diligence Loop sets the strategic direction (the container), while the Description Discernment Loop fills that container with rich, evolving interaction.

By teaching description and discernment as an ongoing practice, we prepare students for sustained, sophisticated collaboration. We are going beyond teaching a technical skill; we are cultivating a new form of intellectual partnership that will shape how they think, create, and solve problems throughout their careers.