"AI Side Hustle for Hospital Administration" - The Day I Used AI the Wrong Way [True Story Vol. 5]

Former hospital administrator'sAIside hustle true story — The realization episode

(Click here for the previous article)

To be honest, until this day, I had been using AI completely the wrong way.
Moreover, I only realized it after I had grilled the AI.

That was when I realized it.

Looking back at the events so far

From Vol. 2toVol. 4, this is what happened.

The AI kept telling me to "create a service page on CrowdWorks." A page that didn't exist anywhere.

No matter how many times I told it that I couldn't find it, it wouldn't admit it. I grilled it, and grilled it, and made it check the user manual, and only then did it finally admit its mistake.

What became clear at that moment was this:

The AI had misunderstood from the beginning that CrowdWorks was a platform where you could list services.

In other words, the AI had been guiding me from the start while confusing it with the mechanism of "Coconala," where workers can create their own service pages.

Those reading this far might think:

"Isn't that the AI's fault?"

I thought so too. But then, I realized something decisive.

A decisive realization

After the grilling was over, I looked back on it calmly.

The AI had misunderstood how CrowdWorks worked. That was a fact.

But I was the one acting exactly as the AI told me.

When it said "create a service page," I searched for it without doubt. When I couldn't find it, at first I thought, "Is my way of searching bad?" I barely considered the possibility that the AI was wrong from the start.

I was the one who was acting by believing the AI's information without verifying it.

I have seen many workplaces where operations are not organized. Mistakes happen. Then, time is taken up dealing with those mistakes. No one thinks it's strange. That's because they are convinced that's just how things are. In hospitals, those mistakes can be a matter of life and death. Registering a different patient with the same name by mistake and performing an unnecessary test. Things like that happen. Because no one notices, I have been the one to improve them.

Yet, I wasn't able to do the same thing for the AI.

"Not verifying" is the most dangerous thing

When I started using AI, I had the recognition that "AI is a convenient tool." But in reality, there were parts where I was acting on the premise that "If the AI says so, it must be correct."

I don't think this is just about me.

AI speaks with confidence. It asserts things. So it's easy to think, "It must be right." But that confidence is not a guarantee of accuracy. AI continues to speak in the same tone even when it is wrong.

In the case of this experience, at the moment it first told me to "create a service page on CrowdWorks," I could have found out in 5 minutes if I had checked it myself. If I had just confirmed what kind of platform CrowdWorks was, this situation wouldn't have happened.

But I didn't check. I acted by believing the AI.

When receiving words from AI, I needed to keep asking myself, "Is this true?"

How to change it specifically

After this experience, I changed how I interact with AI.

----------

Before changing:
AI says something -> Act on it as is

After changing:
AI says something -> Check primary information
　　　　　 -> Verify it myself before acting

----------

Especially regarding things like platform specifications, fees, and rules,"things that exist as facts"I made sure to always check them myself.

What AI is good at is"thinking"and not"researching". For the latest information and specific specifications, I go to official sources, not AI. I realized that this distinction is important.

Looking back at this entire series

To summarize the flow from Vol. 1 to here, it looks like this.

----------

Vol. 1: AIside hustle started
→ Direction found, started moving

Vol. 2: AIchanged what it was saying
→ Listing page not found, incident occurred

Vol. 3: AIinterrogated
→ Made it admit, the tone changed

Vol. 4: AIrevealed its true feelings
→ Finally, the correct information came out

Vol. 5: It was me who was wrong
→ I was acting on AI's word without verifying it

----------

This story, which started with "AI is acting strange," ended with the realization that"I was the one acting strange by trusting AI without question." It's not about AI being bad or AI making a mistake.

AI is a tool I can use right now. But to master it, I needed an attitude of using it critically, rather than believing in it blindly.

What I finally understood on this day

AI certainly made a mistake.

But I was also wrong for acting without verifying it.

It was an event caused by both the AI's error and my own blind trust.

AI is not a magic tool.

Depending on how you engage with it, it can be a useful partner or a useless one.

Next Episode Preview

In the next article, I will write about what it means to "use AI correctly." Based on this experience, I will specifically summarize how to engage with AI to draw out truly useful answers.

My Honest Current Status

------

Earnings: Almost zero
CrowdWorks: Registration complete, waiting for projects
Coconala: Planning to register (next step)
Relationship with AI: Starting to understand it a little

------

I'm not making money yet.

But I am starting to see the right way to interact with AI.

This is my honest situation right now.

(Continues to vol.6 )

👉Past articles in this series are collected in a magazine.

*If you are thinking about a "side hustle with AI," let's go through trial and error together. Send questions to my X DM.*