Writing with Integrity: Why and How I Use ChatGPT as My Co-Author and Safeguard
A behind-the-scenes look at how I use ChatGPT to efficiently write responsible, ethical articles for my Substacks. And a little about my cultural and scientific background.
Preface
Many of my readers have asked why I always mention that my articles are written with help from ChatGPT. So here’s my response—and at the same time, a live demonstration of how I do it: writing efficiently, responsibly, and with joy for my two Substack publications.
See also: An Open Letter to the Informant: Why I Write, and Why I’ll Keep Writing
Why I Always Acknowledge ChatGPT in My Articles
If I had to describe what ChatGPT does for me, I’d call it my Executive Writing Assistant. It's not just a tool or chatbot—it's the equivalent of having a personal editor, researcher, fact-checker, ethics advisor, and formatting expert all in one. It helps me take spontaneous, multilingual voice notes and turn them into structured, ethical, and professional articles that meet the standards expected of a public servant and scientist.
Introduction
There’s an old Soviet joke by the beloved satirist Mikhail Zhvanetsky from Odessa (Ukraine) —the city of my grandparents where I spent many childhood summers— that perfectly captures how I feel when I start writing:
“You should only write when you just can’t hold it in anymore.”
(In Russian, the word "write" is a pun — it can also mean “to pee.”)
That’s exactly how it works for me. When thoughts build up to the point that I can't keep them inside, I let them out—not by sitting down at a desk, but by speaking into my phone. I narrate freely, often switching between Russian and English, depending on which language better expresses the thought. Then ChatGPT helps me shape those spontaneous monologues into a structured and polished article.
That’s how every article I write begins. But why do I always include an acknowledgment of ChatGPT’s help? There are three key reasons: practical expression, ethical protection, and scientific transparency.
1. It helps me express ideas I can’t always write down
English isn’t my first language, and even in my native tongue, complex thoughts often resist being typed out. I also rarely write at a desk. My best ideas come while I’m walking, jogging, or commuting. During those moments, I dictate my thoughts out loud—sometimes in Russian, sometimes in English—letting ideas flow naturally.
ChatGPT processes this narration and turns it into a well-organized article. That’s not just convenient; it’s essential for someone who wants to communicate clearly and consistently, without being slowed down by keyboard or grammar.
2. It helps ensure ethical clarity and protects against misinterpretation
Anyone working for the Government of Canada is bound by the Values and Ethics Code for the Public Sector, which demands we remain non-partisan, truthful, and respectful in all our public communications—even in personal writing shared online.
That’s why I ask ChatGPT to:
Review the text for political neutrality and factual integrity;
Validate key claims using trusted public sources;
Ensure alignment with public service ethical expectations.
This is not just about sounding professional—it’s about protecting public servants from misinterpretation or political backlash. In today’s climate, even well-intended messages can be taken out of context or weaponized by overly zealous partisans from any political direction.
Using ChatGPT as a review tool helps me guard against that. It allows me to communicate honestly and responsibly—while staying well within the ethical boundaries expected of all public servants.
3. It follows the scientific norm of proper acknowledgment
As someone with a background in science, I’m used to acknowledging every tool and contributor—whether it’s software, datasets, or collaborators. To me, ChatGPT is no different. It plays an important role in shaping my writing, and I treat it like any other scholarly instrument: with transparency and credit.
If you’d like to see how I approach technical authorship and acknowledgment in professional work, and also to see what my professional work—as a senior data scientist within the Government of Canada—is about, I invite you to peek into one of my latest reports:
Design and Evaluation of Biometric-enabled Interview Assisting Traveller Screening (a.k.a. Automated Lie Detection) Technology, Defence Research and Development Canada, Contract Report, DRDC-RDDC-2018-C223, November 2018 - PDF (publications.gc.ca)
Collaborative Data Science within Government of Canada: Development of R libraries for common tasks with Open Canada data, Proceedings of Statistics Canada Symposium 2021 (GC Data Conference 2021) - PDF (statcan.gc.ca)
4. The WA command that does it all
The process I follow when writing articles with ChatGPT isn’t as simple as running a single command—it’s iterative, layered, and tailored to my workflow.
First, I tell ChatGPT a lot—sometimes over multiple sessions, often switching between English and Russian—sharing my thoughts as they come to me. Then I usually suggest possible titles and subtitles. After that, I ask ChatGPT to help refine wording, identify missing pieces, find trusted references, and validate any factual claims I’ve made using trusted sources.
Second, I ask it to ensure the entire article aligns with the Values and Ethics Code for the Public Sector (VECPS), maintaining political neutrality and public service integrity.
Lastly, I instruct ChatGPT to count the words, estimate the time spent, and draft the acknowledgment section, as shown at the bottom of this article.
This full process has been defined in a command called WA (“Write Article”), which instructs ChatGPT to do the following:
Validate key claims using trusted sources and populate a list of references. (These links are then verified by me for validity and relevance. If they’re not sufficient, I rerun my WA prompt in Deep Research mode. For simpler articles like this one, the default mode—which relies on built-in knowledge base—is often sufficient.)
Once validated, transcribe everything I told it in a style and wording as close as possible to my original spoken narration, ensuring neutrality and alignment with the Public Servant Code of Values and Ethics.
Generate multiple title and subtitle options, a references section, disclaimer, and acknowledgment section that summarizes my prompt and provides an approximate word count and time based on my narrated input.
This three-stage approach ensures each article is not only an accurate reflection of my voice—but also meets the ethical standards of a public servant and the rigor expected in the scientific community.
References
Disclaimer
This article’s opinions are those of the author and do not represent any institution. It is not intended as legal or medical advice.
Acknowledgment
This article was written with assistance from ChatGPT using the prompt: "Integrate the Zhvanetsky anecdote into my WA-format article as a humorous and personal opener. Preserve the structure and clarity of the original outline, remove personal references, and emphasize the general responsibilities of Canadian public servants."
Based on approximately 1,500 words and 15 minutes of narrated input and collaborative drafting with the author.
ChatGPT was also used to ensure political neutrality, factual accuracy, and alignment with the Public Servant Code of Values and Ethics. Read more about why and how I use ChatGPT to write my Substack articles here.
PS. Videos for your viewing pleasure
One of the many gems from Zhvanetsky (short, humorous bit)
A longer (6 mins) outlook on Soviet reality — with English subtitles (video)