AI That Asks Questions vs. AI That Gives Answers

A professor of teacher education published a piece in Education Week last week that deserves more attention than it will probably receive. Anne Tapp Jaksa, who teaches at Saginaw Valley State University and formerly chaired the American Association of Colleges for Teacher Education, argues something specific and important: AI can technically perform the relational acts of teaching, but doing so erodes the foundation that makes learning stick.

She is not arguing against AI. She explicitly states that AI has legitimate roles in translation, practice generation, grading assistance, and activity suggestions. Her concern is narrower and more useful: when AI performs the relational work of teaching — the noticing, the encouragement, the back-and-forth that Harvard's Center on the Developing Child identifies as essential to brain development — something irreplaceable is lost.

Jaksa is describing a design problem, not a technology problem. And for those of us building AI-assisted learning platforms for Career and Technical Education, her framing is exactly right.

The Design Divide That Determines Outcomes

Most AI entering classrooms right now is optimized for one thing: reducing friction. Give the student the answer faster. Generate the summary. Produce the draft. These tools are efficient and genuinely useful in administrative contexts.

But applied to learning itself, efficiency is the wrong optimization target. When a student receives an answer instantly, the cognitive struggle that builds long-term retention never occurs. When AI produces the summary, the student never develops the reading comprehension required to extract meaning independently. The interaction Jaksa describes — a teacher notices confusion, asks a probing question, the student struggles, the student builds understanding — is not inefficiency. It is the mechanism of learning.

This is not a philosophical objection to AI. It is a design specification. AI built to deliver answers creates dependency. AI built to ask questions creates capability.

Why This Matters More for CTE Than Any Other Sector

Career and Technical Education students face a version of this challenge that is more acute than in traditional academic programs. CTE students are often the first in their families pursuing technical credentials. They face compounding barriers — language gaps, reading level gaps, first-generation student isolation — that answer-delivery AI does not address and may worsen.

Consider two scenarios. An AI system answers the question "how do I calculate a circuit's resistance?" trains the student to retrieve answers from AI. An AI system that responds with "what do you already know about voltage and current — and how might those quantities relate?" trains the student to reason through problems independently. The first student needs the AI present every time. The second student is building the transferable capability that the industry certification exam measures and that employers actually hire for.

This is not a small distinction when CTE programs are accountable to employer partnerships, Perkins funding outcomes, and certification pass rates. The design philosophy of the AI your students use directly affects those numbers.

The Question Every Institution Should Be Asking

When evaluating AI tools for student support, Jaksa's framework suggests a simple diagnostic: does this AI bring students closer to understanding, or deliver understanding to them?

Ask the vendor: what happens when a student asks for the answer directly? An honest answer to that question tells you more about educational outcomes than any feature sheet.

At Core Learning Exchange, our AI platform is built around the Socratic method because the evidence behind guided discovery learning is decades deep. Students who work through questions — who are kept in "question space" longer before resolution — retain material more durably and transfer skills more reliably to new contexts. Our AI platform is designed to do exactly what Jaksa recommends: support human educators rather than replace them, surface mastery data to instructors rather than make decisions for them, and extend the reach of great teaching rather than simulate it.

We are currently enrolling institutions in our research pilot to measure whether this approach produces measurably better CTE outcomes. Because design philosophy, no matter how well-reasoned, should ultimately be backed by data.

Read the Full Article

Anne Tapp Jaksa's opinion piece provides important grounding in developmental psychology and a practical policy framework for evaluating AI in education. It is recommended reading for any administrator currently selecting or evaluating EdTech tools.

Read "AI Can Read to Our Children. That Doesn't Mean It Should." on Education Week

About Core Learning Exchange: We provide turnkey Career and Technical Education (CTE) solutions for grades 6-14, offering 450+ courses from 20+ providers aligned to state standards and industry certifications. Our AI platform uses proven Socratic methodology to develop critical thinking skills through personalized, adaptive learning — deployed in hours via LTI integration with Canvas, Moodle, D2L, and Blackboard.

Related Posts