AI Writing for Instructional Designers: Benefits, Risks, and Best Practices

Recently, on Instructional Designers in Offices Drinking Coffee (IDIODC), we were joined by Connie Malamed, founder of The eLearning Coach, to talk about a question many learning teams are quietly wrestling with:

Your team is probably using AI to write learning content. But is it actually making it better?

As AI writing tools become more common in instructional design workflows, learning teams are experimenting with how AI can support research, outlining, and content creation. The conversation surfaced something deeper than prompt tips or tool comparisons. It challenged us to examine how AI interacts with the cognitive process of writing itself.

We discussed the subtle fingerprints AI leaves in content and how instructional designers can use AI effectively without surrendering their discernment.

If your L&D team is using AI writing tools to accelerate course development inside an authoring and LCMS solution like dominKnow | ONE, this is a conversation worth unpacking.

Writing Is Not Linear. It’s Recursive.

Connie referenced research by Flower and Hayes on the cognitive processes involved in writing. Their model breaks writing into three primary activities:

• ‍Planning
• Translating abstract thought into language
• Reviewing and revising

One key insight that is easy to forget: writing is not linear. It’s recursive.

Writers move constantly between generating ideas, organizing structure, translating thoughts into language, and evaluating clarity. A monitoring process guides these transitions as the writer rethinks ideas and reshapes the message.

For instructional designers, this matters.

When developing structured learning experiences in your content authoring solution, instructional designers are constantly revisiting structure, intent, clarity, and audience alignment.

AI can assist at multiple stages. But unless it is used collaboratively and deliberately, it can skip the essential recursive thinking process that makes learning design effective.

Where AI Helps in the Writing Process

The session explored where AI can support instructional designers without undermining expertise.
‍

1. Idea Generation and Research Support

AI can help instructional designers:

• Surface background information quickly
• Suggest outlines
• Help frame research questions
• Summarize source materials

Connie highlighted research-focused AI tools such as Consensus and Elicit that prioritize academic sources over general web content.

This is important for virtually any learning, but even more critical for instructional designers building compliance training, technical instruction, or evidence-based learning strategies. Directing AI to use materials from well-researched, approved sources can be the difference between a helpful assistant and AI-generated slop.

Bad information does not just affect one course. It can scale across multiple programs.

That is why discernment is essential.

2. Structuring and Organizing Ideas

Many instructional designers use AI to draft outlines before building modules in content authoring and management solutions such as dominKnow | ONE.

This can:

• Accelerate initial structure
• Reveal missing components
• Prompt deeper thinking during revision

Reviewing and critiquing AI-generated outlines often becomes a learning process in itself. Designers understand their material and learners’ needs better and thus must actively evaluate what the AI produced.

If your process or content authoring solution already steers you towards identifying learning objectives early, AI-generated outlines can provide a useful starting framework. The structure can then be refined using your own expertise and knowledge of the audience.

Teams working with structured content repositories may also use these outlines to identify existing learning assets and reuse them. Reuse can then pay off, saving time and improving consistency across learning programs.
‍

3. Editing and Mechanical Support

AI excels at:

• Grammar correction
• Spelling checks
• Clarity suggestions
• Sentence simplification

Used this way, AI functions as a more advanced proofreader.

Of course, this doesn’t eliminate the need for strong writing skills. As noted during the discussion, AI is still an assistant, and sometimes those assistants can make mistakes hear (here) and there.
‍

Using AI to Prepare for SME Interviews

One particularly practical insight from the discussion was using tools like NotebookLM to prepare for subject matter expert interviews.

By:

• Uploading research papers
• Generating summaries
• Creating guiding questions

Instructional designers can enter SME conversations better prepared.

This improves efficiency and reduces the risk of overwhelming SMEs with unfocused discovery questions. It also helps reduce the time required from SMEs’ busy schedules while allowing instructional designers to validate and deepen their understanding during the conversation.

The Fingerprints AI Leaves Behind

One of the most intriguing parts of the session was identifying the subtle patterns that reveal AI-written content.

Overused Sentence Structures

You have likely seen it:

“Instructional design is not about slides. It is about experiences.”

This “It’s not X, it’s Y” structure appears repeatedly in AI-generated writing.

Used once, it can be effective. Repeated throughout a piece, it becomes formulaic and obvious.

Paragraphs That Sound Impressive but Say Nothing

AI often produces text that:

• Uses complex vocabulary
• Feels polished
• Contains little real substance

In learning content, this is dangerous. Training that sounds authoritative but lacks depth undermines learner trust.

This problem can compound quickly in large organizations. Enterprise L&D teams that manage high-volume content production cannot afford surface-level writing. When content scales, weaknesses scale as well.

Emotional Flatness

Another fingerprint is tonal inconsistency.

AI can simulate emotion, but it does not possess lived experience. For reflective learning, scenario writing, or leadership development content, authenticity matters.

Instructional designers bring nuance, context, and audience sensitivity. AI does not.

Content Accuracy

AI can produce sounds accurate, but due to poor sourcing and analysis it may still be incorrect. Even when citations are provided, those citations may not exist, or may not actually support the statements made by the AI agent.

To reduce risks, teams should:

• Break AI use into smaller steps so errors are easier to detect, and they don’t compound
• Direct AI towards trusted sources or known reference material
• Remain active partners in the writing process rather than passive recipients

The Risk: Losing Writing Muscle

One concern raised during the discussion was agency.

If instructional designers outsource the translating stage of writing entirely to AI, what happens to their writing skills?

Writing is a cognitive exercise. It strengthens thinking, clarity, and conceptual synthesis.

In L&D, those skills directly influence:

• Learning objectives
• Scenario realism
• Assessment quality
• SME interview effectiveness

Ultimately, learning content still depends on human judgment to ensure learning impact.

AI should accelerate thinking, not replace it.

AI Inside a Governed Content Ecosystem

The real opportunity for L&D teams is not simply “using AI to write faster.”

It is integrating AI thoughtfully within a governed content ecosystem.

With dominKnow | ONE, teams can:

• Maintain structured, reusable content
• Apply consistent brand and instructional standards
• Control publishing profiles
• Scale updates globally
• Centralize content management

AI can support content generation. But dominKnow | ONE ensures that content is:

• Managed
• Reviewed
• Versioned
• Distributed correctly

That combination protects learning integrity at scale.

For teams interested in structured authoring and governance best practices, the dominKnow Community offers ongoing discussions and practical guidance.

And for instructional design insights from Connie Malamed, her articles and podcasts at The eLearning Coach provide thoughtful perspectives on writing, visual design, and learning strategy.

Discernment Is the Real Skill

Throughout the session, one word surfaced repeatedly: discernment.

Not judgment. Discernment.

Discernment is:

• Evaluating source credibility
• Identifying hallucinated citations
• Recognizing hollow phrasing
• Adjusting tone for your target audience
• Deciding when to rewrite from scratch

AI does not eliminate quality control. It makes it more necessary.

The more AI is used in content production, the more important discernment becomes.

When to Use AI. When to Fight Through the Draft.

Perhaps the most powerful takeaway from the session was this:

Sometimes the struggle of writing is what keeps your thinking sharp.

Use AI to:

• Brainstorm
• Summarize
• Surface research
• Edit mechanics

But consider writing your first draft yourself when:

• You need clarity of thought
• You are developing a nuanced argument
• You are crafting scenario dialogue
• You are shaping learning strategy

The messy first draft is often where real learning design happens.

Final Thoughts for Instructional Designers

AI is here. It is not going away. It is already embedded in authoring tools, research workflows, and content platforms.

The question is not whether instructional designers will use AI.

The question is whether we will use it intentionally.

For L&D teams developing learning at scale with dominKnow | ONE, the combination of:

• Human discernment
• Structured authoring
• Content governance
• Strategic AI use

helps ensure learning content is not just faster to produce, but genuinely better.

Because learning is not about filling space with words.

It is about clarity, accuracy, and meaningful impact.

And that still requires human expertise.

‍

Want to learn more? Watch the replay of our Instructional Design Writing: When to Use AI, When to Resist and sign up to join us for future IDIODC episodes.