The AI Trap: Why Efficiency Is the Wrong Metric for Learning

Integrating AI into education is often framed as a way to clear obstacles and speed up the path to graduation. This view is flawed. By optimizing for task completion, we risk bypassing the friction that makes learning stick. The real competitive advantage for students in 2026 is not using AI to finish work faster, but using it to engage more deeply with the material. Successful learners treat these tools as thought partners rather than shortcuts. For educators and students, the shift from task completion to knowledge mastery is the defining challenge of the next few years. Those who prioritize process over output will possess a lasting advantage that automated workflows cannot replicate.

The Friction-Efficiency Paradox

There is a common belief that we should use AI to reduce friction in the learning process. While this solves the immediate problem of a deadline, it creates a downstream cost: the loss of cognitive development. Justin Serenzia, Executive Director of the Center for Teaching and Learning at The Episcopal Academy, argues that schools often mistake task completion for learning. When students use AI to skip the hard work of synthesis, they are not just saving time; they are eroding their own intellectual foundation.

The systems-thinking approach is to invert the goal. Instead of using AI to remove friction, we should use it to inject it. By using platforms like Viva, which forces students to defend their written work through oral examination, educators can leverage AI to scale the kind of deep, Socratic engagement that was previously impossible in large classes.

> "AI, we can use AI really well to reduce friction in the learning process and to complete tasks and schools oftentimes like require students to complete tasks and then we call it learning but I am not sure how much learning often goes on even predating AI students look for shortcuts to complete tasks. But we can also use AI to inject lots of really good friction along the way."
>
> -- Justin Serenzia

The Thought Partner vs. The Essay Writer

In the college search and application process, the temptation to use AI as a generator is high. Chrissy Serenzia, Senior Associate Director of College Counseling, notes that while half of students are already using these tools, the quality of the output depends entirely on the quality of the input. A student who asks AI to find a college gets a generic list; a student who treats AI as a thought partner to stress-test their preferences gets a strategic advantage.

The danger lies in the black box nature of these tools. When students outsource their voice and judgment to an LLM, they lose the ability to articulate what actually matters to them. The systems-level consequence is that the college admission process becomes a feedback loop of homogenized, AI-generated applications, making it harder for institutions to identify genuine curiosity.

> "AI may be useful as a thought partner but brainstorming questions for brainstorming questions organizing ideas or letting know what matters most on the college experience but it should not replace the student's own voice thinking."
>
> -- Jess Ricker (quoted by Chrissy Serenzia)

Why We Cannot Write the Blueprint

The most non-obvious insight from this discussion is the rejection of the expert model for students. We are in a period of such rapid technological flux that attempting to train students for a specific AI-driven future is a losing game. The system is changing too fast for any static curriculum to keep pace.

Instead, the advantage lies in cultivating agency. Students who learn to vibe code an app or solve community problems using AI are developing a philosopher builder mindset. They are not just learning to use a tool; they are learning to navigate complexity. This creates a lasting moat: while others are waiting for a blueprint of how to use AI correctly, these students are already building the reality they want to inhabit.

> "It is a point in somebody's life that you never know, like we cannot write the blueprint because it is going to change 20,000 more times over the course of the next four or five, 10 years."
>
> -- Justin Serenzia

Key Action Items

• Shift from Output to Process (Immediate): Stop asking "How can I finish this assignment faster?" and start asking "How can I use this tool to stress-test my current understanding?"
• Audit Your Thought Partner (Next 30 Days): When using AI for research or planning, document the prompts you use. If your prompts are generic ("Write an essay about X"), you are outsourcing your thinking. If they are specific ("Challenge my assumption about Y"), you are using the tool as a partner.
• Prioritize Oral Defense (Ongoing): If you are a student, practice explaining your work out loud. If you are an educator, prioritize assessment methods that require students to demonstrate mastery in real-time, bypassing the ability to use AI as a silent ghostwriter.
• Cultivate Play as a Metric (12-18 Months): As AI makes productivity cheaper, human value will shift toward innovation and curiosity. Dedicate time to playing with new tools, building small, low-stakes projects, to build the agency required to navigate the next wave of technological change.
• Ask Better Questions of Institutions (Next College Visit): Move beyond "What is your major?" and ask, "How does the faculty here use AI to deepen classroom inquiry?" or "Are there opportunities to engage with AI as a tool for research, or is it restricted?" This reveals whether the institution is merely reacting to the technology or actively integrating it into their pedagogy.