Emergent scholarship

Course Lessons

Writing assistance

18 min read

Lesson overview

Objective: Apply functional application literacy to writing whilst developing critical evaluation skills for preserving voice—completing the substitution stage of AI literacy development

Summary: This lesson completes the substitution stage. You’ll use AI for bounded writing tasks—drafting support, clarity improvements, alternative phrasings—whilst maintaining the voice that makes your writing yours. Your distinctive voice comes not from what AI generates but from what you choose to keep, how you adapt it, and what you reject as not fitting your writing.

Key habits:

  • Critical selection: AI drafts options, you select what fits your voice and reject what doesn’t
  • Read aloud testing: Read every AI-assisted piece aloud to catch generic patterns your eyes miss
  • Voice retention tracking: Measure how much you revise AI drafts to ensure healthy revision depth (10-30% unchanged is ideal)

The contrast

Easy reading is damn hard writing.

Nathaniel Hawthorne

Dr. James Chen needs to email his research collaborator about a methodological concern. He opens a blank document. Types: “I wanted to discuss the sampling approach.” Deletes it. Too formal. Types: “Hey, quick question about the methodology.” Too casual. Deletes it.

Twenty minutes later, he’s still staring at the opening paragraph. Finally sends something that doesn’t quite say what he meant. Total time: 35 minutes. Still unsatisfied.

Dr. Elena Martinez has a similar email to write. She spends 3 minutes creating a structured prompt asking AI to draft an email explaining a methodological concern to a colleague. AI generates a draft. She then spends 7 minutes revising:

  • Removes “It is important to note that…” (too formulaic)
  • Changes “recent discussions” to “our conversation last Tuesday” (specific context)
  • Replaces “this approach may present challenges” with “I’m worried this sampling method won’t capture seasonal variation” (direct, in her voice)
  • Adds example from their shared pilot study (specific knowledge AI couldn’t have)

She sends it confidently. Total time: 10 minutes. Says exactly what she wanted.

Before we begin

Think about your last difficult writing task. What made it challenging? Where did you get stuck?

AI as writing assistant, not writer

You’ve used AI for content creation (lesson 3) and reading (lesson 4). The same principle applies to writing—AI assists with specific tasks while you maintain control through critical selection.

The crucial distinction: AI as assistant versus AI as writer. AI drafts options, you select critically. AI suggests revisions, you decide what fits. AI generates alternatives, you choose what sounds like you. Your voice, judgement, and intellectual work remain central.

Your distinctive voice comes not from what AI generates but from what you choose to keep, how you adapt it, and what you reject as not fitting your writing. This critical selection is where ethical awareness (maintaining scholarly integrity) and critical evaluation (recognising what’s actually better) intersect.

This lesson completes the substitution stage. You’ll use AI for bounded writing tasks—drafting support, clarity improvements, alternative phrasings—whilst maintaining the voice that makes your writing yours. This is instrumental efficiency: getting unstuck faster, revising more efficiently, whilst preserving what matters most.

Quick reflection

What makes your writing recognisably yours—word choices, sentence rhythm, how you explain ideas?

Voice calibration: Establishing your baseline

Before using AI for writing assistance, you need to calibrate—understanding what “your voice” actually sounds like and what happens when AI changes it.

Calibration insight: This exercise shows you what AI does to your voice. Notice patterns: Does AI make you more formal? Remove specific examples? Add formulaic phrases? This awareness is essential for critical evaluation when using AI for new writing.

Workflow 1: Getting unstuck on drafting

The hardest part of writing is often starting. AI can help overcome blank page paralysis while you maintain control over what you ultimately write.

Faded practice: From observation to independence

Stage 1: Observe expert application

Here’s how an experienced academic uses AI to overcome writer’s block while preserving voice:

[Initial prompt] “Draft an email to my department head explaining that I need to postpone my teaching observation due to a research deadline. Key points: apologise for short notice, explain the deadline is external and inflexible, suggest two alternative dates next week, emphasise I value the observation process. Tone: professional but friendly—we have a good working relationship.”

[AI generates draft]

[Critical revision with annotations]

Original AI: “I hope this email finds you well. I am writing to inform you…” → Revised: “I hope you’re well. I need to ask about rescheduling…” (Less formal opening, matches their relationship)

Original AI: “Due to unforeseen circumstances relating to my research obligations…” → Revised: “The journal wants final revisions by Friday—I didn’t anticipate…” (Specific, not vague)

Original AI: “I would be grateful if we could identify a mutually convenient time…” → Revised: “Would either Tuesday or Thursday morning next week work for you?” (Direct, with specific options)

Notice:

  • Kept: Email structure, professional framework
  • Changed: Every sentence to match voice, added specific details, removed formulaic phrases

Self-explanation

Why did the revision change “unforeseen circumstances” to “journal wants final revisions by Friday”?

Show answer

Specificity replaces vagueness. “Unforeseen circumstances” is generic AI phrasing that could mean anything. The specific deadline shows respect by explaining the actual situation. Your voice emerges from being direct and specific rather than vaguely formal.

Stage 2: Apply to your work

Think of a routine email or paragraph you need to write. Create a structured prompt that gives AI enough context to draft something useful.

When this works well:

  • Routine correspondence with clear information
  • Overcoming blank page paralysis
  • Administrative updates or summaries

When you need to write from scratch:

  • Sensitive communications requiring careful tone
  • Core arguments with your specific examples
  • Anything where relationship context matters deeply

Workflow 2: Clarity improvements

Once you have draft text, AI can identify unclear passages. You then critically evaluate whether suggested improvements preserve precision while increasing clarity.

The approach

Prompt structure:

“Review this paragraph for clarity. Identify sentences that are unclear or unnecessarily complex. For each, explain why it’s problematic and suggest a clearer alternative. [Paste your paragraph]”

Critical evaluation (most important step):

  • Does simpler version lose important nuance?
  • Does it sound like you?
  • What actually improves clarity vs what makes it bland?

Remember: Complexity isn’t automatically bad. Sometimes ideas are genuinely complex and clarity requires precision, not simplification.

Literacy note: Critical evaluation means recognising the difference between necessary complexity (serving precision) and unnecessary complexity (serving no purpose). This is scholarly judgement that AI can’t replace.

Typical time saved: Self-editing a paragraph: 10-15 minutes. With AI highlighting specific issues: 5-7 minutes.

Workflow 3: Generating alternative phrasings

When you’ve used the same phrase repeatedly or can’t find the right wording, AI generates alternatives you select from critically.

Voice preservation exercises

The most important skill for AI-assisted writing is recognising and preserving your distinctive voice. Let’s practice identifying generic patterns versus distinctive writing.

Exercise 1: Spot the generic phrases

Read this paragraph and identify which phrases are generic AI patterns vs distinctive scholarly voice:

“It is important to note that recent research has revealed significant patterns in the data. Furthermore, this analysis suggests that the implications are particularly noteworthy. My interviews with teachers showed three unexpected disruptions in their daily routines. Moreover, these findings indicate substantial variation across different school contexts.”

Mark each phrase:

  • “It is important to note that…” → Generic / Distinctive
  • “recent research has revealed…” → Generic / Distinctive
  • “Furthermore, this analysis suggests…” → Generic / Distinctive
  • “My interviews with teachers showed three unexpected…” → Generic / Distinctive
  • “Moreover, these findings indicate…” → Generic / Distinctive
  • “substantial variation across different school contexts” → Generic / Distinctive

Exercise 2: Revision challenge

Here’s a paragraph written in generic AI style. Revise it to sound like academic writing with voice:

“It is evident that the educational landscape has undergone significant transformation. Furthermore, it can be argued that these changes have had considerable impact on teaching practices. Moreover, recent studies suggest that educators have adapted their approaches accordingly. This development is particularly noteworthy in the context of contemporary pedagogical frameworks.”

Compare to this example revision:

“Teaching has changed fundamentally over the past decade. Thompson’s 2023 study shows teachers now spend 40% more time on assessment than on instruction—a shift that transforms their daily work. My interviews with fifteen secondary teachers reveal they’ve developed three distinct strategies for managing this tension between evaluation and pedagogy.”

What makes the example revision stronger?

  • Specific rather than vague (“40% more time” vs “significant transformation”)
  • Active rather than passive voice (“teachers spend” vs “has undergone”)
  • Ownership and evidence (“My interviews” vs “recent studies suggest”)
  • Concrete over abstract (“assessment vs instruction” vs “contemporary frameworks”)

Decision point: When to use AI for writing

Not every writing task benefits from AI assistance. Let’s practice recognising when AI helps versus when it hinders.

The scenario

You need to write a difficult email to a colleague who missed an important deadline for your joint grant application. This delay means you have only 2 days instead of 2 weeks to finalise the budget section.

What’s your approach?

Pause and reflect

When should you use AI for writing? When should you write entirely yourself?

Decision principle: Use AI for structure and routine content; write sensitive, relationship-dependent, or core intellectual content yourself. Don’t let AI handle tone when relationships matter.

Read aloud practice: The voice preservation test

Reading your writing aloud is the single best test for voice preservation. Let’s practice this systematically.

Literacy note: This read-aloud practice trains your ear to recognise when AI has influenced your voice. With practice, you’ll catch these patterns during writing rather than needing a separate review pass.

Voice retention metric: Measuring revision depth

Voice preservation isn’t just subjective feeling—you can measure how much you revise AI drafts.

Calculate your voice retention score:

For any text where you used AI assistance:

  1. Count total sentences in final version: ___
  2. Count sentences kept completely unchanged from AI draft: ___
  3. Calculate percentage: (unchanged ÷ total) × 100 = ___%

Interpreting your score

10-30% unchanged: Healthy range

  • You’re using AI for structure/drafting support
  • Heavy revision preserves your voice
  • This is effective AI-assisted writing

30-50% unchanged: Moderate voice risk

  • Check carefully by reading aloud
  • Verify phrases sound like you
  • May indicate accepting AI phrasing too readily

Over 50% unchanged: Voice likely compromised

  • Generic AI prose probably dominates
  • Requires more aggressive revision
  • Consider whether AI helped or just produced text you accepted

Under 10% unchanged: AI may not have helped

  • If almost complete rewrite, might have been faster to write yourself
  • AI assistance added overhead rather than value
  • Some tasks don’t benefit from AI drafting

Pause and reflect

Calculate your voice retention score from today’s exercises. What does this tell you about your revision practice?

Activity

Key takeaways

  • AI as assistant requires critical selection: AI as writing assistant means you maintain control over voice, argument, and quality through rigorous critical selection. AI drafts options, you select what works, reject what doesn’t, and revise everything to sound like you. Your distinctive voice comes from what you choose to keep, how you adapt it, and what you reject—not from AI’s initial draft.

  • Voice preservation distinguishes literacy from competence: Many people can get AI to draft text. Fewer develop the critical evaluation skills to recognise what preserves versus compromises their distinctive voice. This requires ethical awareness (maintaining scholarly integrity), critical evaluation (distinguishing genuine improvements from generic alternatives), and contextual judgement (knowing when your voice matters more than efficiency).

  • Read aloud to catch generic patterns: Reading your writing aloud is the single best test for voice preservation. Your ear catches formulaic phrases, awkward rhythm, and generic academic prose that your eyes miss. Make this systematic: read every AI-assisted piece aloud, mark phrases that don’t sound like you, revise them immediately.

  • Substitution works for bounded tasks, not core intellectual work: Substitution-level writing support works for routine correspondence, overcoming blank page paralysis, clarity improvements on existing text, and generating alternative phrasings. It works less well for core arguments requiring your specific examples, sensitive communications needing careful relationship-aware tone, and complex scholarly claims where precision matters more than efficiency.

Your commitment

Pause and reflect

Based on this lesson, what’s one writing task you’ll try with AI assistance this week? How will you preserve your voice? Document this commitment in your Action Journal.

Moving beyond substitution

You’ve completed the substitution stage. Across five lessons, you’ve developed recognition (understanding what AI is), functional application (communicating effectively), applied competence (using AI for content creation, reading, writing), critical evaluation (honest assessment), and beginning taste (pattern recognition about meaningful engagement).

What comes next: Adaptation

The next lessons move beyond substitution to adaptation—where your practice begins reshaping around AI capabilities. You’ll learn to:

  • Use AI not just for bounded tasks but for developing competence in new areas
  • Approach complex problems differently because AI enables new methods
  • Build sustained AI partnerships that accumulate understanding over time
  • Make sophisticated contextual judgements about when to reshape practice

Adaptation requires the functional competence and critical evaluation you’ve built through substitution. You can’t thoughtfully reshape practice without first developing baseline competence. Literacy develops through stages, not leaps.

Resources

  • Mollick, E. & Mollick, L. (2023). Assigning AI: Seven approaches for students, with prompts. SSRN Electronic Journal.
  • Fyfe, P. (2022). How to cheat on your final paper: Assigning AI for student writing. Notices of the AMS, 69(11).
  • Pinker, S. (2014). The sense of style: The thinking person’s guide to writing in the 21st century. Penguin.
  • Sword, H. (2012). Stylish academic writing. Harvard University Press.
  • Silvia, P. (2007). How to write a lot: A practical guide to productive academic writing. APA.
  • Williams, J. & Bizup, J. (2017). Style: Lessons in clarity and grace. Pearson.

Graph View