23 Chapter 14: AI for Writing

“If you want to be a writer, you must do two things above all others: read a lot and write a lot. There’s no way around these two things that I’m aware of, no shortcut.” Stephen King, On Writing: A Memoir of the Craft (p. 145)

Good writing doesn’t begin on the page, it begins with reading. As Francine Prose (2006) argues in Reading Like a Writer, quality writing is deeply rooted in quality reading. Reading with a writer’s eye means not only understanding what is being said but paying close attention to how it’s said: the structure, tone, word choice, and rhythm. These nuances can’t be fully learned through theory alone; they’re absorbed through exposure, practice, and reflection. Stephen King (2000) echoes this in On Writing, where he notes that he reads “seventy or eighty books a year” and finds it hard to believe that anyone who reads little can expect others to appreciate their writing (pp. 145–147).

In this chapter, we shift our focus to the other side of the scholarly equation: writing. Just as AI can support the work of reading and searching the literature, it can also play a powerful role in the writing process, serving as an editor, feedback partner, stylistic assistant, and even creative catalyst. In the following pages, we’ll share how you can strategically use AI to support your research writing while still keeping your thinking, your voice, and your ideas at the center of the process. But before we do that, let us discuss the mindset and philosophy that should guide our use of AI in writing, one that treats these tools as thoughtful collaborators in the creative and intellectual process.

1. Writing as Thinking

One of the core arguments we advance in our philosophy of AI is that we should approach these tools as co-thinking partners, as intelligent collaborators that support our cognitive processes without overtaking them. Nowhere is this partnership more evident than in the act of writing, which, as discussed earlier, is itself a form of thinking. Writing is not merely the transcription of ideas; it is the very terrain where ideas are developed, refined, and brought into clarity. In this context, AI can act as a catalyst for deeper intellectual engagement, amplifying our ability to think critically and creatively.

Thinking, at its essence, is mediated through language. Whether it’s the silent voice of inner thought or the externalized expression of speech and writing, language is the scaffold on which cognition is built. The richer and more nuanced our linguistic repertoire, the more complex and sophisticated our thinking becomes. While we do not subscribe entirely to the doctrine of linguistic determinism, we do believe that language shapes thought, not absolutely, but profoundly. The extent to which we can conceptualize, explore, and communicate our ideas is inseparable from our command of language. If we cannot express an idea clearly, its communicative power is diminished, and its intellectual potential is compromised.

This is why mastery of language (i.e., articulateness, fluency, eloquence) is foundational to cognitive expression. And this is where generative AI steps in as a transformative force. When we use tools like ChatGPT, Claude, or Gemini to assist with the language component of thinking, we offload some of the cognitive burden that often gets entangled in surface-level concerns: grammar, syntax, word choice, sentence flow. This reduction in cognitive load liberates mental energy, allowing us to focus more intensely on the substance of our ideas. In other words, AI doesn’t think for us, it makes space for us to think better.

Imagine how freeing it is to simply write, to pour your thoughts onto the page without being hindered by the fear of making linguistic mistakes. You no longer need to halt your flow to second-guess a sentence’s structure or worry over a vocabulary choice. That freedom, that creative fluency, is where AI shows its true power; it removes the friction that slows us down. This is what we mean when we say AI can be a catalyst for deeper thinking.

Now consider this same scenario from the perspective of a second-language writer attempting to publish academic work in English, a language they did not grow up speaking. The act of writing becomes doubly demanding. Not only must they wrestle with complex ideas, but they must also encode those ideas in a language that may still feel foreign or imperfectly mastered. A large portion of their cognitive effort is spent on ‘languaging’ rather than thinking.

This is a critical reason why many second-language academics are at a systemic disadvantage when competing with native speakers in scholarly publishing. While all academic writers must learn the conventions of academic discourse (since it is a code that must be learned, not naturally acquired) native speakers start from a position of linguistic familiarity. For second-language writers, however, language often becomes a barrier to expression rather than a bridge.

With the advent of large language models, this playing field is beginning to level. These tools, when used wisely, allow non-native speakers to refine their prose, clarify their arguments, and meet the rhetorical standards of academic writing with far less strain. In this sense, AI is democratizing access to strong writing by extending the scaffolding of language mastery to anyone willing to learn how to use it effectively. And that, in itself, is a profound shift in how we think, write, and participate in knowledge production.

Building on this understanding of AI as a cognitive and linguistic aid, we’d like to share how we personally integrate these tools into our writing practice, including the writing of this book. Don’t get us wrong, AI tools, at least in their current state, cannot replicate the depth, nuance, and soul of human creative writing. They lack the lived experience, intuition, and emotional resonance that true creativity demands. However, what they can do, and do remarkably well, is enhance human creativity. We genuinely believe that anyone not leveraging the power of AI to assist with their writing is missing out on one of the most transformative tools of our time. Whether you’re a seasoned author with several publications under your belt or a novice scholar still navigating the contours of academic writing, generative AI offers an incredible opportunity to tap into deeper layers of thinking by shifting the burden of linguistic precision to the machine.

Our writing strategy begins with what we call the “free-flow draft.” This is a raw, unfiltered expression of thought, our ideas, arguments, and intuitions laid bare, without concern for grammar, sentence structure, or stylistic polish. To protect this creative freedom, we would disable spelling and grammar alerts in our editors; we don’t want red underlines interrupting the rhythm of our thoughts. We would write as freely as a bird soaring across an open sky. It is a liberating process, thinking through writing in its purest form. Once this draft is complete and our ideas are safely captured, we then begin the process of proofreading and refining. At this point, we enlist the help of AI, primarily ChatGPT or Claude, although Gemini’s latest 2.5 version is quite impressive as well.

At the heart of this process is a deeply collaborative approach to writing, what we described earlier as working with AI as a co-thinker. These tools don’t generate our content; they elevate it. They help us enhance clarity, polish tone, and improve coherence. Importantly, this strategy ensures that every idea is authentically ours. While we occasionally use AI for early-stage ideation (e.g., exploring prompts, questions, or directions during brainstorming) for academic and research writing, we prefer to ground our ideas in the literature of our field and our own background knowledge. These ideas, shaped through reading, reflection, and accumulated experience, form the foundation of meaningful academic contribution. They are the fruit of understanding and the springboard for further insight, innovation, and scholarly dialogue.

In this sense, we treat AI tools more like professional editors than co-authors. Once the intellectual labor of drafting is done, we would hand the baton to our AI assistant for stylistic refinement. We would ask it to help with smooth transitions, vary sentence structure, and elevate the overall quality of expression. This phase of the process is where AI chatbots like ChatGPT and Claude truly help.

Now, let’s discuss some of the helpful ways to use AI chatbots in your writing!

2. AI Chatbots

In the next sections, we’ll outline practical strategies for using AI chatbots specifically ChatGPT and Claude to elevate your academic writing. While we focus primarily on these two tools, the same principles apply to other AI chatbots you may be using. We’ll also introduce a range of additional AI-powered tools that go beyond chat-based interfaces, tools designed to support different stages of the writing process, from idea organization to language refinement.

2.1. Work in Projects

One of the most powerful features available to paid users of ChatGPT and Claude is the Projects workspace. As discussed in the previous chapter on AI for reading, this feature allows you to create a dedicated, persistent space for a specific task, whether it’s reading a long academic article, analyzing a dataset, or, as in our case here, editing and refining a full-length research paper or book manuscript.

When working on an extensive writing project, we always prefer to set it up as a Project. Unlike the regular chat interface, Projects maintain a consistent and extended context window. This means that the AI remembers your ongoing work, understands the broader structure and purpose of your text, and can return to prior inputs to maintain coherence across sessions. This continuity is incredibly valuable when writing something as intricate and layered as an academic paper.

To get the most out of this feature, we recommend setting clear custom instructions at the outset. In the instruction panel, you can explicitly assign the role of professional editor to the chatbot. Ask it to focus solely on improving language (e.g., enhancing style, refining tone, tightening coherence, and polishing flow) without altering your original ideas or arguments. This ensures that the AI acts as a linguistic assistant, not a co-author (check out chapter 1 to learn more about Projects and how to set them up).

Here’s a sample custom instruction you might use:

“You are a professional academic editor. Your job is to improve the clarity, coherence, and style of my writing without changing any of the original ideas or introducing new content. Focus on grammar, sentence structure, flow, and academic tone. Do not remove or reframe my arguments just help me express them more effectively.”

You can always use the main chat interface for smaller tasks or quick feedback. But for long-form writing, Projects offer a more structured, efficient, and context-aware environment that supports deeper, ongoing collaboration between you and the AI.

2.2. Editing your Draft with AI

Once you’ve established your project workspace and assigned your AI tool the role of a professional editor, the next step is to begin the actual editing process. Our key recommendation here is to work in small, manageable sections, ideally a few paragraphs at a time. While it might be tempting to paste the entire draft and ask for a full edit in one go, in practice, We’ve found this method to be less effective. The more focused and contained the input, the more precise and relevant the AI’s output will be.

Editing in chunks allows you to monitor and preserve the integrity of your ideas. It also gives you better control over tone, consistency, and style across different parts of the text. This is especially crucial in academic writing, where logical progression, precision, and voice matter a great deal.

Even after the AI provides its edited version, it’s important to remember: your job as the writer isn’t over. The draft it returns should be treated as a polished base, not a final product. You’ll often want to revise it further to ensure that your unique tone and intent remain intact, especially when expressing complex or nuanced arguments.

To guide the AI more effectively during editing, always set a clear role and provide specific instructions. For academic writing, here are several sample prompts you can use to get the best results:

Sample prompts for section-based editing

1. General Academic Editing

“You are a professional academic editor. Please improve the clarity, style, grammar, and flow of the following section. Maintain the original meaning and structure of the arguments. Do not introduce new content, focus strictly on language and coherence. [Insert section]”

2. Transitions and Structure

“You are a professional academic editor. Suggest improvements to transitions between paragraphs and ensure the logical flow of ideas. Recommend heading hierarchy if applicable and identify places where structural improvements could help readability. [Insert section]”

3. Citation and Referencing Consistency

“You are a professional academic editor. Check that the citations in this section are properly formatted and consistent with APA/MLA/Chicago style (insert your preferred style). Identify any missing references and suggest how to integrate them smoothly into the narrative. [Insert section]”

4. Terminology and Argument Consistency

“You are a professional academic editor. Ensure consistent use of terminology and conceptual phrasing throughout this section. Check that key terms are used clearly and that the logical flow of arguments is coherent and well-developed. [Insert section]”

5. Conciseness and Precision

“You are a professional academic editor. Please revise this section for clarity and conciseness. Remove wordiness, unnecessary repetition, or vague expressions while preserving the depth of the ideas. [Insert section]”

Crafting detailed and targeted prompts like these will help the AI understand exactly what kind of support you need. Over time, you’ll develop a rhythm for working with AI as a trusted editorial assistant, one that enhances the academic quality of your writing while preserving your own intellectual voice and contribution.

2.3. Generating feedback

Another important way to use AI chatbots like ChatGPT to help in the writing process is by getting it to provide you critical feedback on your manuscript. Just as a peer reviewer or writing mentor might offer constructive critique, these tools can provide a fresh perspective on your work, helping you spot gaps in reasoning, refine your claims, and anticipate possible counterarguments.

For example, you might ask it to assess the clarity and strength of your argument, suggest areas where the logic could be tightened, or highlight parts that feel underdeveloped or unsupported. You can also ask the model to evaluate the internal coherence of your ideas, do they build naturally on one another? Are there any leaps in reasoning or missing connections?

Another powerful way to use AI in this stage is to request counterarguments. When you ask the model to challenge your position or offer alternative perspectives, you expose your ideas to critical pressure which is an essential part of scholarly thinking. This can be especially useful when preparing for peer review or anticipating objections in academic debate.

Here are a few sample prompts researchers can use to guide this process:

Sample prompts for Feedback and critical review:

1. General Feedback

“You are a professional academic editor and critical reader. Please read the following section and provide constructive feedback on the strength of the argument, clarity of ideas, and logical flow. Identify any areas that are vague, underdeveloped, or repetitive. [Insert section].”

2. Check for Missing Support

“You are a professional academic reader. Please review the following argument and identify where additional evidence, citations, or examples might be needed to strengthen the claims. [Insert paragraph or argument]”

3. Suggest Counterarguments

“You are a critical peer reviewer. Please identify possible counterarguments or alternative perspectives to the position presented in this section. Frame them in a way that challenges the logic but remains academically constructive. Insert argument”

4. Critique Structure and Development

“You are a professional academic editor. Analyze the structure and progression of the following argument. Are the points logically sequenced? Are transitions effective? Suggest how the section could be reorganized for better flow and coherence. [Insert section]”

5. Balance and Tone

“You are a professional academic editor. Evaluate whether the following section maintains an objective and balanced tone. Flag any areas that appear too biased, assertive without support, or emotionally charged for academic writing. [Insert content]”

Using generative AI in this way turns it from a language assistant into a thinking partner, one that helps you engage more critically with your own work. While it doesn’t replace peer feedback, it can certainly supplement it, especially during early stages when you’re still shaping your arguments. With clear prompts and a collaborative mindset, ChatGPT and Claude can push your thinking forward and help you refine your research with greater clarity, rigor, and depth.

3. AI Writing Tools

ChatGPT and Claude are definitely powerful collaborators for editing, refining, and critically engaging with your drafts, but they are just one part of a broader ecosystem of AI tools available to researchers today. Beyond chat-based interfaces, there exists a growing suite of specialized platforms designed to support different aspects of the academic writing process, from organizing ideas and managing references to checking grammar, detecting tone, and even formatting your manuscripts. In the next section, we introduce several of these tools and explain how they can be strategically integrated into your workflow to enhance the clarity, structure, and overall impact of your research writing.

3.1 Jenni

Jenni is a fully AI-optimized platform specifically designed to support the research and writing process from start to finish. Jenni offers an integrated environment that combines text generation, citation management, and real-time AI interaction, all within a distraction-free workspace tailored for academics.

One of Jenni’s most helpful features is its AI-powered autocomplete, which intelligently suggests sentences and ideas based on what you’ve already written. This feature is especially useful for overcoming writer’s block or generating momentum during the early drafting stages. You can also highlight any sentence or passage and right-click to choose from options like edit, improve, or regenerate making it easy to refine your text as you write.

Another interesting feature is Jenni’s in-text citation generator. As you’re writing, simply highlight an argument or claim, right-click, and select Citation. Jenni will instantly suggest relevant academic sources from its connected databases allowing you to choose a study or article to cite. The platform then formats the in-text citation automatically according to your chosen style guide: APA , MLA , Chicago, Harvard, or IEEE. This feature streamlines what is often one of the most tedious parts of academic writing.

Jenni also includes an embedded chat assistant called AskJenni, which opens a sidebar interface where you can engage with the AI while working on your document. You can ask it specific questions about your draft, request help with rewriting sections, or seek clarification on a concept. AskJenni pulls insights from both your personal library and the web to generate helpful, context-aware responses.

Speaking of the library, Jenni allows you to upload multiple PDFs and research papers and create a personalized archive of sources. You can chat directly with your uploaded documents, summarize them, extract key points, explain complex concepts, and even compare insights across papers.

3.2. Paperpal

Paperpal is another AI-powered tool to help you with your academic writing. It offers a suite of intelligent featurss to help improve your writing style, clarity, and academic tone. As you begin writing in Paperpal’s built-in editor, the platform provides context-aware suggestions to complete sentences, rephrase awkward phrasing, and elevate your language. You can also highlight a sentence and choose from various enhancement options such as making it sound more academic, paraphrasing it for originality, or trimming it for conciseness. For more targeted improvements, you can also highlight individual words to receive custom synonym suggestions thus helping you find the most precise vocabulary for your discipline.

Paperpal also offers a collection of writing templates that support a range of academic tasks including abstracts, titles, personal statements, essays, keywords, study highlights, and more. To use a template, simply select the one that fits your task, fill in the required fields, and let Paperpal generate a tailored draft based on your input.

Beyond writing assistance, Paperpal offers plagiarism detection, translation tools, and access to Editage, a professional editing service staffed by human experts for researchers seeking more advanced review and language polishing. For those who prefer working in Microsoft Word, Paperpal provides a convenient Word add-in that enables you to access its full suite of services directly within your existing documents.

 3.3. AI-Powered Word Editors

After exploring how to use AI chatbots like ChatGPT and Claude, along with specialized AI-powered tools such as Jenni and Paperpal, let’s now shift our focus to another important element of the writing process: your word editor. No matter how powerful  AI tools are, you’ll still need a robust, flexible writing environment to bring everything together. AI is an incredible enhancer, but it works best when embedded in or combined with a solid editor. The good news is that many of our most trusted, long-standing editors like Microsoft Word and Google Docs have evolved to integrate AI capabilities directly into their platforms. These tools, once seen as “old-school,” are now smarter, faster, and more supportive of the academic writing process than ever before.

Let’s take a closer look at some of these editors and how they can support your work.

3.3.1. Microsoft Word

Microsoft Word remains one of our go-to platforms for extensive academic writing projects like dissertations, theses, and book chapters. It’s reliable, feature-rich, and widely supported across institutions and platforms. But what makes Word even more indispensable today is how it has evolved, particularly through its integration with AI.

If you have a Copilot Pro subscription, you can bring the full power of generative AI right into the Word editor. With Copilot, you can ask the AI to generate summaries of your chapters, assist with writer’s block, complete unfinished sentences, or even suggest rephrasings. One feature we find particularly useful is “Visualize as Table”. You simply highlight a section of text and convert it into a well-formatted table in seconds. This can be a major time-saver when dealing with structured information or data-heavy sections of your writing.

Another strength of Microsoft Word is its support for add-ins (i.e., mini extensions that add more features and functionalities to the editor). These are easily accessible through the Microsoft App Store, and many of them are particularly useful for student researchers. Here are a few we  regularly recommend:

• Grammarly for Microsoft Word: While Word already includes built-in grammar and spell-check tools, Grammarly goes further. It offers real-time suggestions for improving sentence clarity, eliminating unnecessary words, rephrasing awkward constructions, and even adjusting tone. It’s especially helpful when you’re stuck or unsure how to express a complex idea clearly.
• ProWritingAid: Like Grammarly, this tool offers spelling and grammar corrections but takes it a step further with in-depth style reports. It helps you spot and remove clichés, identify paragraph structure issues, highlight vague language, and ensure consistency in things like hyphenation and capitalization. It’s a great tool for fine-tuning your academic writing.
• QuillBot: An AI-powered tool for paraphrasing and summarizing. QuillBot helps rephrase sentences, improve sentence flow, find more precise vocabulary, and even condense large chunks of text, all without changing the core meaning of your content.
• Power Thesaurus: This handy add-in integrates seamlessly into Word giving you quick access to synonyms and antonyms. It’s perfect for enhancing vocabulary variety and avoiding repetition.

These add-ins make Word a dynamic writing environment especially when paired with conversational AI tools like Copilot, ChatGPT or Claude. While you could technically switch between Word and a separate chatbot interface, using AI directly within Word via Copilot or integrated add-ins creates a much smoother and more efficient workflow.

3.3.2 Google Docs

With the recent integration of AI features into the Google ecosystem, Docs has become a much more powerful and capable writing tool, one that now stands as a serious alternative to Word. One of the most helpful additions is the “Help Me Write” feature. This tool allows you to input a prompt and have the AI generate a range of text, from outlines and summaries to paragraphs and full sections. While it’s useful for brainstorming and idea generation, we find its greatest value in editing and proofreading. Simply highlight a section of your document, right-click, and select “Refine selected text.” From there, you can choose various AI-driven actions like changing the tone, summarizing, elaborating, shortening, bulletizing, or even applying a custom instruction tailored to your editing needs.

Like Microsoft Word with its Copilot integration, Google Docs is now powered by Gemini, Google’s flagship generative AI engine. You’ll find the “Ask Gemini” icon in the top right corner of your document. Clicking it opens a side panel where you can engage in a live chat with Gemini while working on your paper. You can ask it to brainstorm ideas, generate an outline, summarize sections of the document, or input your own prompts for specific tasks. This real-time AI support, embedded directly in your document, makes the writing process more interactive and efficient.

Beyond AI features, one of the enduring strengths of Google Docs is its seamless integration with the Google Drive ecosystem. If you use tools like Sheets, Slides, or Drawings, you can easily pull in tables, charts, and visuals directly into your writing. And with offline editing, you can continue working even without an internet connection, just keep in mind that changes won’t sync until you’re back online.

With these capabilities, Google Docs has come a long way from being a simple online word processor. It’s now a robust writing environment that supports every stage of the academic writing process from drafting and outlining to editing and collaboration.

3.3.3. Scrivener

If you’re looking for a more professional and sophisticated program tailored specifically for long-form academic writing, Scrivener is one of the best options available. It offers a host of powerful features that can significantly streamline the thesis or dissertation writing process.

Scrivener embeds a unique organization-first approach. You can easily break your manuscript into chapters, sections, or sub-sections, and work on each part individually without losing sight of the bigger picture. One of its interesting features is Scrivenings mode, which allows you to stitch your sections together and view them as a continuous document, perfect for reviewing and editing your work in context.

Another unique organizational feature is the virtual index card system. Each section of your project is linked to an index card on a digital corkboard. Rearranging your structure is as simple as dragging and dropping these cards. Scrivener automatically reorders your manuscript accordingly. It’s a visual and intuitive way to structure complex projects.

The Outliner is another practical feature. It gives you a bird’s-eye view of your entire manuscript and lets you jump to any section instantly. You can reorganize, annotate, and plan your structure with ease. And if you ever need to compare chapters or sections side-by-side, you can split the editor and have up to four documents open within the same project window.

Scrivener also includes a full-featured editor that rivals Word in terms of formatting and styling capabilities. Its Focus Mode is particularly helpful. It hides everything else on screen, allowing you to concentrate fully on your writing without distractions. The platform also supports essential academic features like tracking changes, word count goals, custom labels and color coding, footnotes, and reference management.

As for sharing options, Scrivener lets you export your manuscript in various formats including Microsoft Word, OpenOffice, RTF, PDF, and more. Scrivener is available for macOS, Windows, and even iOS so you can work from wherever you are. While it may take a bit of time to get used to, Scrivener is an excellent tool for academic writers who need power, flexibility, and structure in their writing process.

3.3.4. Notion

Although often associated with note-taking, Notion is far more than a digital notebook. It’s a flexible, all-in-one productivity platform that’s increasingly being adopted by researchers, students, and writers for managing and producing complex writing projects including theses, books, and research papers.

At the core of Notion’s writing workflow are Pages, which function as digital workspaces. You can write your entire thesis in one Page or break it down into separate Pages for each chapter or section. Each Page supports rich formatting and allows you to insert images, tables, links, and more. You can also collaborate in real-time, inviting peers, supervisors, or co-authors to edit or comment directly in your workspace.

One of the best features for academic writers is the ability to turn Pages into custom templates. This is incredibly useful when working on structured documents like research papers or grant proposals where you follow a repeated format.

Notion also includes Notion AI, an integrated AI assistant built directly into the platform. You can activate it at any time to generate outlines, summarize sections, rewrite text, translate passages, or even create visual elements like tables or diagrams. What makes Notion AI particularly effective is how well it integrates with your existing content. You can ask it questions about your writing, have it find specific items across your Notion workspace, or generate content in response to your ongoing project.

Additionally, Notion’s visual organization tools like databases, kanban boards, calendars, and timelines make it easy to manage tasks, track deadlines, and map out your research workflow. Everything is neatly housed within a single, highly customizable space.

Whether you’re using it to organize your reading notes, draft sections of your thesis, or manage your research timeline, Notion offers a flexible and highly personalized environment to support your academic writing process.

Conclusion

Academic writing has always been a deeply cognitive, creative, and at times, solitary endeavor. But we are now witnessing a shift, one where AI is not replacing the writer, but walking alongside us as a collaborative partner.

In this chapter, we explored how generative AI, particularly tools like ChatGPT, Claude, and a growing range of AI-powered platforms, can enhance the academic writing process without compromising intellectual integrity. From drafting and editing to receiving feedback and managing citations, these tools serve as co-thinkers that support, extend, and refine our ideas.

We shared our personal strategies for using AI in a way that safeguards the originality of thought while improving expression, clarity, and structure. We also highlighted the importance of writing environments (e.g., Word, Google Docs, Scrivener, Notion) that now integrate AI features, creating seamless workflows where generative assistance meets the practical needs of long-form academic writing. These editors, combined with smart AI integration, provide the infrastructure researchers need to write more productively and purposefully.