🤫 AI's Big Secret: What 700M People Use It For (It's Not Work)

Today, AI (Artificial Intelligence) is everywhere. The most famous tool is probably ChatGPT. A lot of people are using it. A big new study looked at exactly how 700 million people use ChatGPT.

The researchers wanted to find out what people really do with this tool. The results are very interesting. They show us how AI is changing our daily lives.

In this article, we will look at the biggest findings from the study. Some things might surprise you. It turns out, AI is not just for work.

Finding 1: AI Is Used More For Personal Life Than for Work

This is maybe the most surprising thing. When ChatGPT first came out, many people thought it was a tool for work. People thought it would help programmers write code, help marketers write ads, or help bosses write emails.

But the study shows a different picture.

Only 27% of all chats with ChatGPT are for work.

This means the other 73% is for personal things. People are using AI for their daily lives, hobbies, learning, and personal interests. This number has grown a lot. Just one year ago, personal messages were only 53%.

This shows a big change. This is not because new, casual users are joining. It is because current users are changing how they use it. They are finding more ways to use ChatGPT in their personal lives.

Why is this happening?

There are a few possible reasons:

1. AI is like a friend who knows everything: It is easy to use and always available. You can ask it anything at 3 AM. You don't need to worry about bothering someone.

2. No fear of being judged: Sometimes we have "silly" questions. We might feel shy to ask other people. For example, you might want to ask, "What is inflation, exactly?" or "Explain a movie to me that everyone is talking about." Asking AI is very easy.

3. Worry at work: Some companies do not like employees using AI. They worry about data safety or that employees will become lazy. So, people might feel more comfortable using it at home for personal things.

So, what are those 73% of personal users doing? The chart from the study shows the top three topics:

• Seeking Information: This is using ChatGPT like Google. But it's different. Instead of getting a list of links, you get one direct answer. It can understand what you mean. For example, you can ask, "Find me a fast, healthy dinner recipe with no chicken."

• Writing: Helping write emails to friends, writing social media posts, or even writing poems.

• Practical Guidance: These are "how-to" questions. For example: "How do I fix a bike tire?" or "How do I plan a 3-day trip to Da Lat?"

Finding 2: Writing Is the #1 Use Case At Work

Even though work is only 27% of all use, inside that 27%, there is a clear winner.

40% of all work tasks are about WRITING.

This is not strange. ChatGPT is very good with language.

But here is the interesting part: two-thirds (about 66%) of those writing requests are editing or changing text that the user gives. Only one-third is asking it to write something new from the start.

What does this mean?

It means people don't trust AI to write everything for them. Most of us know what "AI writing" sounds like. It often sounds very general, robotic, and has no feeling (some people call it "AI slop").

So, instead of asking, "Write an email for me," people are using it in a smarter way. They write a draft first, then ask AI to make it better.

Added Value: How People Use AI to Edit (Prompt Examples)

A "prompt" is the instruction you give to the AI. Here are detailed examples of how people use AI to edit, not to write from new:

Example Prompt 1: Make text sound more professional

• Bad Prompt (asking to write new): "Write an email to my boss saying I am sick."

• Good Prompt (asking to edit):

"Here is my email. Can you help me make it sound more professional and polite?

My draft: 'Hey boss, I feel very tired. Can't come to work today. See you tomorrow.'"

• Result: The AI will rewrite it, for example: "Dear [Boss's Name], I am writing to let you know that I am feeling unwell today, [Date], and will not be able to come to the office. I will check emails if I can and hope to be back tomorrow. Thank you for your understanding."

Example Prompt 2: Make text shorter

• Bad Prompt: "Write about the benefits of exercise."

• Good Prompt: 

"This paragraph is too long. Can you help me make it shorter, around 50 words, but keep the main idea?

My draft: 'Exercise is a very important thing that everyone should try to do every day. It not only helps you feel better in your body, but it is also really good for your mind and helps you feel less stressed. Also, when you exercise, your heart beats faster and that is good for your heart. It can also help you sleep better at night.'"

• Result: The AI will summarize: "Exercise is key for health. It improves your body, reduces stress, boosts heart health, and helps you sleep better. Try to be active daily."

Example Prompt 3: Check grammar and mistakes

• Good Prompt:

"Can you check the grammar and spelling in this text? It is for my blog.

My draft: 'Travel is one of my hobby. I has been to many place, like Thailand and Singapore. I think people should travel more because it open your mind.'"

• Result: The AI will fix the mistakes and might suggest better words.

Other Work Tasks (Not Writing)

So, what about the other 60% of work tasks?

• Programming (Coding): Surprisingly, the study shows only 4.2% of messages are about programming. This number is very low. Many people thought ChatGPT was a tool for programmers. The reason might be that professional programmers are using special AI tools, like GitHub Copilot, instead of the normal ChatGPT.

  

• Brainstorming (Getting Ideas): People use it to find new ideas.

  • Example Prompt: "Give me 10 ideas to advertise my small coffee shop on Instagram."

    
    

• Summarizing: This is very useful.

  • Example Prompt: "I have a 20-page report. Can you read it and give me 5 bullet points that summarize the main ideas? [paste the report text here]"

    
    

Finding 3: The "Worrying" And "Strange" Uses

The study also found some uses that make us think.

1. 10% is for Teaching and Learning

Many people, especially students, are using ChatGPT as a personal tutor.

• The good part: It is great. It is patient. You can ask it to explain hard ideas.

  • Example Prompt (Learning): "I don't understand 'photosynthesis'. Can you explain it step-by-step with very simple words, like I am 10 years old?"

    
    

  • Example Prompt (Language Learning): "I am learning English. What is the difference between 'since' and 'for'? Give me 5 examples for each word."

    
    

• The "worrying" part (The Hallucination Problem):

  This is a big problem. An AI "hallucination" is when the AI makes up information but says it with confidence, like it is a fact.

  • Example: You ask, "Who invented avocado toast?" ChatGPT might say, "It was John Avocado in 1952 in California." (This is not true).

    

  • This is very dangerous for students. If they trust the AI 100% and do not check the information, they will learn the wrong things.

2. 2% is for Relationship Advice

People are asking ChatGPT about personal problems. They ask about boyfriends, girlfriends, family, and friends.

• Why? Again, the AI does not judge. You can tell it your secrets and not feel silly. It is always there to "listen."

• Why is it "strange"? AI does not have feelings. It does not understand love, sadness, or jealousy. It just looks at data from millions of texts on the internet. Its advice can be very general and not really good for complex human feelings. It might not be a good idea to rely on AI for your relationship problems.

3. 2% is just Small Talk

People are saying things like, "Hi, how are you?" or "I am sad today, tell me a joke."

This shows that some people might be lonely. They are treating AI like a friend, not just a tool. As technology gets better, the line between a tool and a friend might get blurry.

4. People are "Asking" more than "Doing"

The study found a cool change. The use is split into:

• Asking: 51.6% (Finding advice, decisions, "What should I do?")

• Doing: 34.6% (Write this email, summarize this text)

• Expressing: 13.8% (Just chatting, sharing feelings)

This means people are using AI more like a consultant or an advisor instead of a worker. They want AI to help them think and make decisions. This leads to a big risk, which we will talk about at the end.

Finding 4: AI Is Changing Our Skills

One of the most important findings is that AI affects beginners and experts in very different ways.

1. AI Helps Beginners (Novices) a Lot

A "novice" is a person who is new to something.

If you have never made a website, you can ask ChatGPT to give you the basic HTML and CSS code. You can go from 0% to 80% of the work very fast.

This is a great thing. It helps more people try new things without feeling scared. It lowers the barrier (makes it easier) to learn a new skill.

Example Prompt (Beginner): 

"I want to make a very simple website for my bakery. Can you give me the HTML and CSS code for a homepage with a title 'The Sweet Bakery', one paragraph description, and one picture?"

The Result:

<!DOCTYPE html>
<html lang="en">
<head>
  <meta charset="UTF-8">
  <meta name="viewport" content="width=device-width, initial-scale=1.0">
  <title>The Sweet Bakery</title>
  <style>
    body {
      font-family: Arial, sans-serif;
      background-color: #fff8f0;
      color: #333;
      text-align: center;
      margin: 0;
      padding: 0;
    }

    header {
      background-color: #f4b183;
      padding: 20px 0;
    }

    h1 {
      margin: 0;
      color: white;
    }

    main {
      padding: 40px 20px;
    }

    p {
      max-width: 600px;
      margin: 0 auto 20px;
      font-size: 1.1em;
    }

    img {
      width: 80%;
      max-width: 400px;
      border-radius: 10px;
      box-shadow: 0 4px 8px rgba(0,0,0,0.1);
    }
  </style>
</head>
<body>
  <header>
    <h1>The Sweet Bakery</h1>
  </header>

  <main>
    <p>Welcome to The Sweet Bakery! We bake fresh bread, cakes, and pastries every day using the finest ingredients and lots of love. Come taste the sweetness!</p>
    <img src="bakery.jpg" alt="Delicious pastries and bread">
  </main>
</body>
</html>

2. AI Can Annoy Experts

An "expert" is a person with a lot of skill and experience.

An expert already knows the basic 80% of the work. They need help with the last 20% – the hardest part, the part that makes something great.

This is where AI often fails.

It can do the 80% job well, but it often gets the important 20% wrong. Experts call this "knowledge dilution."

• Example: A professional writer asks AI, "Write a description of a sad beach, in the style of Hemingway."

• The AI might write something with short sentences, but it will not have the deep feeling or the real "style" of Hemingway.

• The expert ends up spending a lot of time fixing what the AI wrote. Sometimes, it takes more time than writing it themselves.

The lesson: AI is a great assistant for beginners, but it is not a replacement for experts. The real value is still in the last 20% that humans do.

Finding 5: Who Is Using AI?

The study also looked at who is sending these 2.2 trillion messages.

1. Half of All Messages Come from Gen Z

Gen Z are people born between about 1997 and 2012. They grew up with smartphones and the internet.

Some people worry about this. They worry that young people will become too dependent on AI. They will not learn how to write, how to think, or how to solve problems by themselves. It is like we are entering a "fast fashion era of knowledge work."

But we need a balanced view.

• The bad side: If you ask AI to write your whole essay, you learn nothing. You are "cheating" yourself.

  • Bad Prompt (Lazy): "Write me a 1000-word essay about the causes of Climate Change."

• The good side: It is different if you use AI as a tool to learn.

  • Good Prompt (Learning): "I have to write an essay on the causes of Climate Change. Can you help me make an outline with 3 main points? I will write the paragraphs myself."

    
    

  • Good Prompt (Checking): "Here is my essay. Can you read it and tell me if my argument is logical? Did I miss any important points?"

    
    

The problem is not the tool. The problem is how we use it.

2. The AI Gender Gap Has Closed

When ChatGPT first came out, about 80% of users seemed to be male (the study guessed this based on names).

But now, that number has dropped to 48%. This means the number of male and female users is almost equal.

This is good news. It shows this tool is useful for everyone.

The study also found a small difference:

• Users with female names tend to use AI more for writing and practical guidance.

• Users with male names tend to use it more for technical help (code) and seeking information.

These are just general observations, not rules. It shows that different people find different value in the same tool.

Conclusion: What Does All This Mean For Us?

We are still in the very early days of this technology. But this study gives us some big clues about the future.

Here is what we can learn:

1. Personal lives will change a lot: AI will make us more knowledgeable. We can learn how to fix things or cook new foods. But there is a risk. If we talk to AI more than we talk to people, we might become lonelier. We need to be careful to stay connected to real humans.

2. 'AI Literacy' (AI Skill) is Most Important: In the future, what you remember might not be as important as how you ask AI. Knowing how to ask good questions (prompting), how to check information, and when not to use AI will be the most valuable skills.

3. Editing is the new writing skill: As we saw, most people are using AI to edit. The human's job is changing from "writer" to "editor." We are the final check. We are the ones who add feeling and judgment.

4. Beware of "Decision Laundering": This is the biggest risk. "Decision Laundering" is when a person uses AI to make a hard decision, and then blames the AI if the decision is wrong.

   • Example: A manager asks AI, "Based on performance data, should I fire employee A or employee B?" The AI gives an answer. The manager fires employee A and says, "The AI decided."

   • This is very dangerous. AI does not have morals. It does not understand the human situation. We must always be responsible for our big decisions.

This study shows that ChatGPT has become a deep, personal tool. It is not just a work machine. It is a tutor, an advisor, a friend (sometimes), and an editing assistant.

Our job now is to learn to use it wisely.

If you are interested in other topics and how AI is transforming different aspects of our lives or even in making money using AI with more detailed, step-by-step guidance, you can find our other articles here:

• Stop Using ChatGPT As An "Answer Machine"! Do THIS Instead

• Stop Building n8n AI Agents Manually! THIS Does It For You*

• Fully Detailed & Powerful Instruction That Drive Custom GPTs/ Projects/ Gems*

• Forget Film School! THIS Is The Future Of AI Video Creation!*

• Master AI Prompting: Get Pro Results From ChatGPT & Gemini

  *indicates a premium content, if any