Conversation with Google Gemini

I was in an online discusssion recently about the potential implications of generative AI as a tool for supporting professional practice, in this case teaching. I asked a question in the discussion (see below). After the discussion ended I used it as a prompt for Google Gemini. At this stage, I was still in my ‘hands off, I don’t want to get involved in AI mode’ although, through following various feeds, I am aware of the problems and dilemmas that generative AI can bring to the role of educators (the very issues that have made me wary of getting too deeply into AI).

Here’s the question and below is Gemini’s reply which is both thought provoking (i.e. it’s pretty good!) and worrying (couldn’t have come up with that in less than a minute!). It’s an impressive reponse to a moderately elaborate  question. Finally I have succumbed! I am now beguiled by these mysterious beasts.

If ‘AI’ can be ‘trained’ to reproduce realistic conversations, facsimiles of novels, essays, mathematical proofs, images, videos, etc. etc., and if, in this way, AI weakens the centrality of people (you and me) in creating these things, with the consequence that for a great deal of cultural activity the role of people is in some way downgraded or sidelined, then does this not also affect, and sideline, the pivotal role of teachers in planning, constructing, and guiding teaching and learning?

Gemini’s Reply:

(For reference here’s a link to the original reply: https://share.gemini.google/wxhQIGFlpwzV)

The short answer is yes, it undeniably destabilizes the traditional role of the teacher—for the exact same reason it destabilizes the author, the artist, and the programmer.

When a technology automates the creation of artifacts (essays, lesson plans, problem sets, rubrics, explanations), any profession whose primary value was bound up in producing those artifacts experiences a crisis of purpose.

However, the nature of this “sidelining” in education is distinct from other cultural domains because of what teaching actually is at its core.

1. The Dislocation of “Teacher as Content Provider & Planner”

Historically, a significant portion of a teacher’s authority and labor rested on two pillars:

• Curriculum Construction: Designing unit plans, scaffolding assignments, generating practice materials, and writing rubrics.
• Knowledge Delivery: Explaining complex concepts, answering foundational questions, and evaluating student output (grading).

Generative models now perform all of these tasks with remarkable velocity. An AI can instantly draft a differentiated 6-week unit plan on Hamlet, produce tiered math problems for varying skill levels, or provide step-by-step diagnostic feedback on a student’s essay.

If teaching is defined primarily as the generation of instructional content and the execution of assessment pipelines, then the teacher’s pivotal role is actively downgraded. The teacher risks becoming a mere middle manager—a human prompt engineer or an administrative monitor enforcing interaction with machine-generated curricula.

2. The Devaluation of Student Artifacts

The shift goes deeper than teacher workload; it transforms the basic contract of learning.

For centuries, education relied on the proxy: “If you can produce the artifact (the essay, the proof, the project), you have acquired the capability.”

Because AI trivializes the production of the artifact, that proxy breaks. When the artifact is decoupled from the cognitive labor of the student, the teacher’s role as an evaluator of finished work becomes largely obsolete or untrustworthy. Grading a take-home essay is no longer evaluating a mind; it is often evaluating an algorithm or an editing process.

3. Where the Pivot Must Shift: The Resistance of the Relational

This raises a crucial question: Is the artifact the point, or is the process the point?

In art and literature, the sidelining of human creators happens because consumers often care only about the final product (the image, the track, the text). But education is not a product-delivery system; it is a process of human transformation.

If teachers are sidelined, it is usually because education systems have treated them as content delivery mechanisms rather than relational mentors. Where AI cannot easily supplant the teacher is in the territory that technology historically ignores:

• Curating Productive Friction: AI minimizes friction, offering instant answers. Learning, however, requires cognitive struggle. The teacher’s role shifts from making things clear to knowing when to make things difficult—introducing boundary conditions, counterexamples, and real-world stakes that force genuine thinking.
• Evaluating the Live Mind: As static homework artifacts lose diagnostic value, teaching must return to Socratic oral defense, real-time problem-solving, physical laboratory work, and collaborative discussion—spaces where human presence and live reasoning are observed directly.
• Social and Epistemic Trust: Motivation in learning is heavily social. Students rarely push through deep intellectual discomfort for a software interface alone; they do it within a community, spurred by human recognition, accountability, and mentorship.

The Core PradOX

The danger is not that AI is incapable of teaching skills, but that educational institutions—driven by cost, scale, and efficiency—will mistake the automated production of instructional materials for the act of education itself.

If education is reduced to data transfer and artifact verification, teachers will indeed be sidelined into technicians supervising software. Preserving the pivotal role of the teacher requires explicitly redefining teaching away from content production and back toward the human practice of intellectual formation.

* END *