GenaI as co-author and more importantly as "Devil's Advocate"

In a companion post on context stacking, I came across an idea that stayed with me — and I wanted to explore it further.

This piece isn’t just about the final blog produced, but about the process behind creating it using generative AI (specifically, Claude.ai). Rather than using AI as a writing shortcut, I used it as a thinking partner — one that could challenge my assumptions, test my reasoning, and help strengthen the argument before anything was finalised.

What emerged was a structured workflow (shown below) that others can adopt when using AI to improve the rigour, not just their output. And it all starts with setting the context and the audience and telling the generative AI to pick it apart.

The process:

Before using Generative AI: Two drafts were produced, and the second draft went through this final process, as described by claude.ai here.

Here's the workflow we followed:

1. Critical Reading and Initial Diagnosis

We started with a close reading of your original draft, analysing what was working well structurally and argumentatively, and identifying the key weaknesses an academic reviewer would likely challenge — particularly the flipped classroom compliance problem you raised directly.

2. Devil's Advocate Questioning — Round One

Rather than moving straight to revision, we used structured devil's advocate questioning across four pressure points: the transferability of the central claim, the wrong mental model problem, academic displacement, and the equity argument. Each question was designed to find where the argument was exposed rather than where it was strong.

3. Iterative Stress-Testing — Multiple Rounds

Crucially, we didn't accept your first answers. Each response you gave was pushed harder in the next round — your answers generated new, more specific questions. This produced several iterations where the argument either strengthened genuinely or revealed honest limitations. Notably, when you challenged my assertion about academic workload lacking evidential basis, that exchange itself modelled the critical thinking the piece advocates.

4. Position Clarification on Key Tensions

Through this process several important clarifications emerged that weren't in the original draft: the champion fragility problem, the coordinated misconception risk, lived experience as embodied presence rather than purely knowledge, and most significantly the honest narrowing of your target audience from "all academics" to "student-centred module leaders with genuine autonomy."

5. Honest Scoping Decision

The pivotal moment was agreeing that the piece was most honestly a proof of concept argument for early adopters rather than a universal prescription — and recognising that this resolved the central contradiction between your equity argument and your actual audience.

6. Structural Redesign Before Drafting

Before writing a single word of the revised piece, we agreed a revised section architecture that mapped the new argument logic — ensuring the structure served the argument rather than retrofitting the argument into the original structure.

7. Drafting with Referenced Integration

Only then did we draft, integrating Biggs and Tang and Freeman et al. as gaps the conversation had specifically identified, and producing a piece that honestly names its scope in the subtitle rather than burying the limitation.

---

The meta-point worth noting

This workflow is essentially the context stacking methodology your piece advocates — applied to academic writing rather than lecture preparation. We built a model, stress-tested it against challenging questions, identified where it broke down, refined it, and only then committed it to final form. The process mirrored the argument

The bigger takeaway

This process highlights a more powerful way to use generative AI. Instead of treating it as a tool for faster writing, it can be used to improve how we think. By asking it to question, challenge, and probe our ideas, we can produce work that is more robust, more honest, and ultimately more useful. On a more personal note, it pushed me hard to think what I was actually thinking and meaning - it didn't give me the answers.

The inverted questioning and lots of multiple rounds were the key strengths of the approach at each stage/round. I was asked some insightful but critical questions. I did want it to give me the answers but push me to address (and face) critical questions.

All opinions in this blog are the Author's and should not in any way be seen as reflecting the views of any organisation the Author has any association with. Twitter @scottturneruon