Can you proudly call AI-generated materials "your own"? What a 40-something AI promotion lead realized on the front lines
Hello everyone, this is Ponzu.
In our AI promotion team, we each brought in project proposals. Using AI is already a standard practice in this day and age.
However, when we laid the materials out side-by-side, something became clearly visible.
Materials that incorporated personal thought versus those that did not. The difference was easier to see than I had imagined.
Today, I am writing about what is necessary to make materials "your own" when creating them with AI.
Our AI promotion team decided that everyone would bring their own opinions and project ideas regarding future activities. We had one week. Each person thought about it in their own way and compiled it into a document. When we opened the lid, there was a greater difference in how we used AI than I had expected.
Materials that were supposed to be streamlined by AI lacked actual thought
A week later, team members brought in their respective materials.
In this day and age, I believe using AI is a given. Using it as a sounding board, brainstorming structures together, or checking for typos—there is no problem with that kind of usage.
However, when we opened the lid, there were materials where the AI output had simply been pasted into PowerPoint. The structure and layout were exactly as the AI had produced them. There was no trace of the person's own words anywhere.
Another person who caught my attention was someone who submitted in Word. I think part of it was that they weren't good at PowerPoint. But honestly, I suspect another reason was that they tried to make it with AI but found it troublesome to make adjustments, and with Word, they could just output it as-is and call it finished.
Both used AI. But neither incorporated their own thinking. The moment the materials were laid out, that difference was clearly visible.
How to spot materials dumped by AI. Two points a 40-something manager noticed
So, how can you tell if a document lacks "incorporated thought"? There are two points I noticed.
Point 1: There are unnecessarily many supplementary explanations in parentheses
When you have AI create materials, it carefully inserts explanations for terms. It's kind, if you want to call it that, but such explanations aren't needed in internal company documents. The readers are members of the same team. When basic terminology is explained to people who already have the prerequisite knowledge, it actually feels out of place.
Point 2: It contains content that doesn't match the current issues
For example, re-organizing information that the team has already shared, or listing general theories that aren't what should be discussed right now. Since AI cannot fully understand context, this kind of output tends to happen if the instructions are vague.
Actually, I have had similar output when I had AI create materials in the past. It is precisely because I remember correcting it by saying, "No, that's not it," that I was able to notice it the moment I saw these materials.
In other words, if you use AI output as-is, the content might "look the part," but it becomes a document that doesn't fit the audience or the situation. Whether or not you can notice this might be the dividing line between whether or not you have incorporated your own thinking.
3 steps to make AI-generated materials "your own"
So, how can you turn AI output into "your own materials"? When I organize what I usually do, it breaks down into three steps.
Step 1: Insert your own thinking
Before dumping everything on the AI, you need to articulate what you want to convey. In my case, I use the AI as a sounding board to organize my opinions and get help with the structure. The important thing here is not to use the structure the AI provides as-is. You must judge for yourself: "Is this order okay?" or "Is this item really necessary?" I believe this process is the first step toward injecting your own thinking into the material.
Step 2: Select and discard information
AI tries to provide as much information as possible in response to what it is asked. However, the amount of information the reader of the material needs is limited. Keep only what the other party wants to know and cut the rest. This "judgment to cut" is something only humans can do for now.
Step 3: Add your own touch
AI creates materials using general expressions, but it cannot replicate your own words or the context within your company. By adding your own phrasing and the premises shared within your team, it finally becomes "your own material."
All three of these are mundane tasks. But the moment you skip them, the material reverts to being "something made by AI."
4 reasons why I still make PowerPoint slides myself
In addition to the three steps mentioned above, there is one more thing I am particular about: I make my PowerPoint slides myself.
There are four reasons for this.
Reason 1: Establishing my position within the company
As an AI lead, being thought of as someone whose materials are easy to understand leads to trust. If I just presented materials made by AI, I wouldn't be able to build this position.
Reason 2: Organizing my thoughts
In the process of building slides one by one, my own thoughts become organized. Thinking, "Will this order convey my point?" or "Is this single slide really necessary?" should itself be training for my thinking process.
Reason 3: To speak in my own words during presentations
Because I made the materials myself, my own intentions are embedded in every single slide. That's why I don't get stuck for words when explaining them. If it were just materials dumped on me, I wouldn't even know what I wanted to say myself.
Reason 4: Maintaining PowerPoint skills
I think this is surprisingly important. Because I have the skill to create slides myself, I can judge the output provided by AI, saying things like "This is wrong" or "This layout is better." Without those skills, I wouldn't even have a standard to evaluate AI output.
That said, I don't believe everything should be done manually. With NotebookLM, you can even create PowerPoint-like slides depending on your prompts. That is precisely why I believe just as Sakamoto Ryoma used a pistol despite being a master swordsman, it is important to choose your tools only after you have acquired the necessary skills.
The first step to improving work efficiency with AI is to "delete the unnecessary parts"
Having read this far, some of you might be thinking, "Honestly, 3 steps or 4 reasons, that's a hassle." I get it. I wasn't like this from the start either.
So, if you are someone who is about to try creating materials with AI, please try just one thing first.
Delete the unnecessary parts from the text the AI has generated.
Don't add; delete. For example, try looking at these things:
Unnecessary half-width symbols
Three consecutive sentence endings (e.g., "masu. masu. masu.")
Unnecessary supplementary explanations in parentheses
Content that doesn't fit the current phase
Excessive decorative language
Just doing this will significantly change the impression of your materials. This is because the act of deleting always involves the thought process of judging "what is necessary and what is not."
You don't need to redo everything from scratch all at once. First, look at the AI's output and try deleting just one part where you think, "Is this really needed?" I believe that is the first step toward making AI output "your own material."
Summary of "Why AI efficiency doesn't lead to 'your own materials'"
Creating materials with AI is no longer difficult. What is difficult is making them "your own materials."
Incorporate your own thinking. Select and discard information. Add your own personal touch. And start by deleting the unnecessary parts.
In my previous article, I wrote that I "hadn't cultivated a culture of how to use it." As a continuation of that, this article is my attempt to organize how to specifically go about it.
I am still halfway there myself, but I hope to gradually spread the idea of "creating together."
With that, I'll take my leave.