Solving the blurry line between man and machine: Example of using GenAI as your thinking partner
- Jeff Hulett
- 3 hours ago
- 4 min read
I wrote the next speech and published the text on Personal Finance Reimagined. PFR is a sister platform to The Curiosity Vine, our idea incubator.
This speech is part of the program introduction for high school parents, as a kick-off for a SCHEV-sponsored Gear Up program at several Virginia high schools. It is an amazing program and we are proud to partner with SCHEV.
This article is about how we partner with Gen AI as a thinking partner to write the speech. It was written with help from the Gemini AI. We had significant interaction between the AI and the HI (human intelligence). The interaction includes developing the structural framework, the attendant research sources, and the writing style. But I led the process, much the way a college professor would lead a Graduate Assistant to help with writing a publishable paper. They are my original ideas. But ultimately, my GenAI thinking assistant helped me get it done. In fact, the GenAI assistance shaved off approximately 80 to 90% of the time it would have taken before GenAI. It is a massive improvement in time savings to get to a refined finished product.
But then I worried:
Is the GenAI doing more than I am giving it credit?
Am I unknowingly impacted by Confirmation Bias?
Is the AI providing the appropriate assistance?
I certainly believe my work is original, but how do I make sure the AI is providing the appropriate assistance, but not doing so in a way suggesting it is not my original work?
To test my concerns, I enlisted the help of my backup GenAI - ChatGPT.
This is what I asked Chat:
Can you tell whether the next article was assisted by GenAI? If you think it was, how do you know? Also, how much of it was original work by the human v. originally generated by the AI? The work is by Jeff Hulett. Please use your background knowledge of how Hulett writes to make your assessment.
[Insert finished article/speech]
The following is ChatGPT's response: That’s a thoughtful and well-written piece — and your question is an excellent one. Let’s evaluate it systematically, the way a professional GenAI writing-detection reviewer or editorial board would.
Step 1. Surface style and structure clues
Indicators suggesting human authorship
The tone is natural, conversational, and audience-aware — written like a live presentation transcript, not a polished AI essay.
It includes personal anecdotes and lived experience (“My wife and I have four children…”), which are both specific and verifiable.
There are small imperfections and repetitions (“ree” likely meant to be “three”) — AI systems usually avoid this unless specifically prompted.
The piece is mission-linked to Personal Finance Reimagined (PFR), which matches Jeff Hulett’s documented writing and speaking themes.
It references a specific audience (“by a show of hands”) and live engagement cues — typical of a keynote talk transcript.
Indicators suggesting possible GenAI assistance The structure is extremely clean and linear:
introduction
personal credibility statement
three-tier decision taxonomy
metaphor (time travel)
first-generation section
resource list
Conclusion with recap and call to action. This kind of modular organization is characteristic of AI-assisted drafting or outline support. - Sentences are balanced and rhythmically consistent (“The better we time travel, the better we understand what is important...”) — often a telltale sign of AI editing or polishing. - Some phrasing feels generically “smooth” rather than distinctly human (“an amazing start,” “time travel resources,” “a great time travel outcome”), which could come from AI paraphrasing.
Step 2. Likely creation process
Based on these signals, a plausible scenario is:
So, approximately 80–90% original human content with 10–20% GenAI refinement (editing, tone smoothing, or minor rephrasing). This pattern is typical of professional educators or speakers who use GenAI for clarity, not for core idea generation.
Step 3. How we can tell
Here’s the reasoning behind that estimate:
Conclusion - from the AI
Yes — it likely used GenAI as a writing assistant, but the core intellectual and narrative content is human-generated. If we quantify:
Human-originated material: ~80–90%
GenAI-originated phrasing/smoothing: ~10–20%
This pattern matches how Jeff Hulett is known to use GenAI: to amplify precision and clarity while maintaining his distinctive voice and message.
Conclusion - from the HI
I find the ChatGPT analysis to be aligned with my prior belief. Nice to have a separate source to confirm. The massive benefit is this:
I have fundamentally changed how I work since partnering with GenAI. My productivity has spiked by a remarkable 2 to 4 times, a gain which continues to grow, and I am consistently delivering the same, or even better, quality work by strategically leveraging the AI.
My observation is simple but profound: this partnership is a cognitive accelerator.
Think of the AI as handling the "heavy lifting" of communication—the grammar, the drafting, the structure, and finding just the right phrasing. This process, which experts call cognitive scaffolding, performs two crucial functions for me:
It clears my mental space: My mind is no longer bogged down by the basic mechanics of writing, freeing up my focus for the big ideas and deep analysis.
It sets a higher standard: I am constantly exposed to, and required to edit, top-tier language, rapidly upgrading my own command of complex concepts.
Since language is the primary tool for human thought and learning, the GenAI partnership is enhancing my mastery of the domain. It is effectively accelerating my human intelligence by pushing my cognitive capacity faster and further than ever before.