AI Risks? Watch Out for These! Points to Note When Using AI - AI Column for Groomers by the CHOKIP Engineering Team, Vol. 1

Hello! This is the CHOKIP engineering team💻
As a new initiative, we are bringing you an article from our engineering team!
We hope to share some behind-the-scenes stories about the development of CHOKIP someday, but
this time, for salon owners and groomers who think "AI sounds interesting but difficult," we would like to share somefirst stepsanduseful informationin a series of articles.
We hope you enjoy reading them!😊

Are you thinking, "What can AI actually do?" or "It sounds difficult"?

To put it simply, AI is likea friend who is happy to give advice and knows a lot of things. For example, just like asking a chef friend for dinner ideas...

Question

AI's Answer

That is one way you can use it🍳
If we look at how to use it for pet salon work...🐶

Question

AI's Answer

What do you think?
Because the instructions in the question were more specific than the cooking example, the answer is also more specific.
There are techniques for how to give instructions and ask questions to AI like this☝️
I will share these techniques on another occasion!

In actual salon work scenarios, there are many cases where the AI's answer is not perfect, such as "weekends are not available" or "Moka-chan is not fluffy, she is curly."
The key is to "not take AI answers at face value," "not use them as-is," but to have a conversation a few times to get advice until you reach an answer close to what you want, and thenit is a very important tip to arrange that text before using it.

Let's prepare to actually use AI

"But doesn't AI cost money?" Many people might be thinking that.
You can actually use it for free for trial purposes or low-frequency use✨
There are several famous services that use AI via chat, but this time we will introduce how to use ChatGPT from OpenAI in the United States, which is the most used in the world and officially announced over 900 million weekly active users as of February 2026.
I plan to talk about AIs other than ChatGPT on another occasion.

1. First, access it
https://chatgpt.com/ Please try accessing this from your computer or smartphone browser.
👇 I think a screen like this will be displayed.

Just by writing text in the input area in the middle and sending it, the AI will give you an answer💡 AI thinks and provides an answer on the spot every time, so it does not necessarily mean the same answer for the same question💻

2. Create a login account
How was it? ChatGPT is easy to use just by accessing it, right?
However, I recommend registering an account rather than using it without logging in☝️
This is because by creating an account, you can check past history and resume conversations✨ Furthermore, by using the same account, you can share the same history between your computer and smartphone app.
And creating an account is also important when avoiding the"AI risks" mentioned later.
To create an account, proceed to "Sign up for free" in the top right.
Then, I think a screen like this will be displayed👇.

If you have a Google or Apple account, you can use that to log in (create an account).
If you use a phone number or email address, enter it along with a password on the screen after pressing "Continue," and you will receive a confirmation notification via SMS or email, so you can create an account by following the instructions💡

3. Log in and try it out
Once you log in, you can use AI by sending text to ChatGPT just like before!
Please consult and ask AI about various things and enjoy your AI life!

...is what I would like to say, but there are points to note when using AI⚠️
Please be sure to check the next"What are the risks of AI? Watch out for these! Points to note when using AI" before using it.

What are the risks of AI? Watch out for these! Points to note when using AI

"AI is convenient but can be dangerous," "I tried using it, but it wasn't as convenient as everyone says."
You may often hear negative opinions like these about AI.

AI is dangerous. It's not very convenient...
These opinions are by no means wrong.
However, by properly understanding the "dangerous parts" and "why you felt it wasn't convenient," you can make your salon work easier, increase sales, and make your life more convenient and enjoyable 🐶

For example,gasolineis the driving force that moves vehicles and allows us to "transport large quantities of people and goods, over long distances, quickly."It is convenient,but unfortunately,it also has dangerous aspects that can lead to accidents.In fact, AI is said to be an industrial revolution greater than gasoline (oil) or electricity, and by properly understanding and using its dangerous aspects, there are many points where it is just as convenient as gasoline.
The introduction has been long, but this time I will tell you about two things:
"Information Leakage"and "Plausible Lies."

<Points to Note Regarding Information Leakage>
AI does not know all the knowledge in the world.
It is designed to answer questions and instructions from users by "learning" information from the internet, dictionaries, papers, etc., in advance.
The content of our conversations with AIusing chat is actually sometimes used for learningto evolve the AI.
As an extreme example, if you search for restaurants near your home in your daily conversations, it might learn your home information.
If you search for real estate purchases, your asset status might be learned.
If you provide credit card information, it might answer the question, "Do you know Mr./Ms. XX's credit card information?"
As you can see, while "learning" has a good side that contributes to AI performance, it also leads to the risk of leaking the user's personal or confidential information.
In ChatGPT, the setting to use chat interactions for learning is ON by default. Turning this setting OFF is the first step to reducing the danger.
I will tell you how to set it up!

① Open the ChatGPT settings screen
When you select your name in the bottom left of the ChatGPT screen, a menu will appear, so select "Settings".

② Disable learning settings
From the items on the left side of the screen that appears next, select "Data Controls" -> "Improve the model for everyone".

Finally,turn the switch to the right of "Improve the model for everyone" to OFF as shown in the image below, and select execute.

With this, your chat interactions with AI will no longer be used for learning.
However, even though it will no longer be learned, just in case, you should absolutely never enter confidential information such as credit card numbers or passwords used on websites, information that can identify individuals, or customer information for pet salons.

<Plausible Lies>
The answers AI gives you may actually be wrong.
That's right. It "tells lies"!?
However, the AI is not lying on purpose or out of malice. I mentioned earlier that AI uses "learned" knowledge.
If the internet information or papers used for learning are wrong, it mistakes them for correct information.
Humans also increase their knowledge by searching the internet or talking with acquaintances, but that can also be wrong. The textbook used for learning was wrong.
Also, depending on the way of thinking, there may not always be "one correct answer."

For example, let's think about "cooking."
Depending on whether a person likes or dislikes rolled omelets, the evaluation or impression of "delicious" or "bad" changes.
The way of making it and preferences vary by region, such as sweeter in Kanto and stronger dashi in Kansai... 🥚
AI uses data from various positions and situations to provide the answer that "the AI thinks" is most suitable.
Therefore, it is important to supplement the position and situation, such as "Tell me how to make rolled omelets for people in Tokyo."
Furthermore, AI may answer things it doesn't know by "pretending to know."
It tries hard to derive "I think this is the answer" from the knowledge it has learned. Therefore, "the AI tried hard to give me an answer, but it was full of nonsense" is a common occurrence for AI users.
If you continue the conversation by digging deeper into what it answered first, or if you press it on small contradictions, it will easily admit, "I'm sorry. The previous answer was wrong" 💦

The important thing is not to take the AI's answer at face value, but to research it yourself or add your own interpretation to create the final answer.
I showed you an example at the beginning of "having AI think of a reply for a cancellation," but if you copy and send the AI's answer to a customer without checking it, it could turn into an unexpected response and cause trouble.
You might feel it's a hassle, but if you understand the dangers and shortcomings of AI and deal with it properly, I think you will find AI surprisingly convenient and reliable!
Please try to challenge yourself, even if it's just a little bit 😊

If you have any other information you would like to know about AI, please feel free to send a message to the "CHOKIP Official LINE Account".

*This article is information as of the date of publication on note.
*The examples of Q&A with AI posted in this article are images to make it easy to understand, not actual AI answers.

If you want to know more about CHOKIP, please apply for the "CHOKIP Introduction Video".

Click here to apply for CHOKIP materials ✨

Click here to register as a friend on the "CHOKIP LINE Official Account" where you can check advantageous information such as campaign information 🪄