ChatGPT Making You Stupid 🤖🧠 Learn English Through Story Level 2 Graded Reader 🎧

Añadido: 2026-05-16

[00:00:00]Imagine the year 2035.

[00:00:03]Life is very different. Computers that can think, called AI, are everywhere. In offices, AI writes emails and makes presentations.

[00:00:14]It helps with reports. At home, AI creates music and even movies. Students don't need to study in the old way. They learn how to use AI tools for their schoolwork. Everything you need seems to be easy to get.

[00:00:32]You just need to ask the AI.

[00:00:35]In the past, this future seemed like a children's cartoon. But now, it feels very real. AI technology is in almost everything we use today. This makes us think about something important. Are we starting to use our own brains less?

[00:00:52]Will we stop trying to solve problems by ourselves? Maybe AI will make our brains softer over the next 10 years.

[00:01:02]Will we become unable to think well for ourselves? This is a big question. It is the main idea of this story. AI can do many amazing things, but we need to think about the price of using it so much. We need to ask, is AI making us less smart?

[00:01:22]Two, a familiar failing. This feeling of changing how we use our minds is not totally new.

[00:01:31]We can look at older, simpler tools to see a pattern. Think about using a GPS to find your way.

[00:01:39]A study in 2020 looked at people who used GPS a lot. It found that this strong use could make their memory for places weaker.

[00:01:50]Using a map app is easy and fast. But this convenience can sometimes have a hidden cost to our skills.

[00:01:59]Other technologies before AI also showed this.

[00:02:04]Think about calculators.

[00:02:06]For many years, using calculators for simple math problems changed how students learned basic arithmetic.

[00:02:14]Autocorrect tools in writing programs also changed things. Studies have shown they can affect students' ability to spell words correctly or use the right punctuation.

[00:02:28]These past examples are important.

[00:02:32]They show that when we start using tools to do a task for us, our own skills for that task can become less strong.

[00:02:41]Calculators helped with math, but some students practiced basic adding and subtracting less.

[00:02:48]Autocorrect helped with typing, but some students paid less attention to spelling.

[00:02:55]These tools were simpler than today's AI, but the effect on how we use our brains is similar.

[00:03:03]They set a kind of example for our situation now. They show that using tools a lot can change our own abilities.

[00:03:12]It's a familiar feeling when we think about how AI is changing things today.

[00:03:18]Three, the lazy brain. Our brains are wired to find efficient ways to do things. This natural tendency is sometimes called cognitive offloading.

[00:03:32]It means we use external tools, resources, or systems to help reduce the amount of mental effort we need to use for a task.

[00:03:42]For example, using a simple checklist for shopping is a form of cognitive offloading.

[00:03:50]With the rise of advanced we now have the ability to offload more complex cognitive tasks that used to require significant mental work.

[00:04:02]A notable study involving over 600 participants from various backgrounds investigated how frequent use of AI impacts critical thinking skills.

[00:04:16]Critical thinking involves analyzing information and making reasoned judgments.

[00:04:23]The study discovered that individuals who used AI tools often were more likely to engage in cognitive offloading.

[00:04:34]They tended to rely heavily on the technology for tasks like problem-solving and decision-making, rather than thinking through problems themselves.

[00:04:45]The findings showed that over time, participants with high reliance on AI demonstrated a reduced ability to critically evaluate information and develop detailed, well-supported conclusions on their own.

[00:05:04]This finding connects to the experience of a professor named David Rafo.

[00:05:10]He noticed a sudden improvement in the writing quality of some students during the pandemic.

[00:05:16]However, upon investigation, he realized that it was not the students' actual writing skills that had improved.

[00:05:26]Instead, they were using AI tools to generate or refine their text.

[00:05:33]Professor Rafo noted that the tools were doing the writing work, not the students building their own abilities.

[00:05:42]This highlights the difference between using a tool to achieve a result, and genuinely developing a skill.

[00:05:50]Our mental and cognitive abilities are often compared to muscles in the body.

[00:05:57]Like physical muscles, mental abilities need regular use and challenge to remain strong and function well.

[00:06:06]If a physical muscle is not used, it can become weak. This is known as atrophy.

[00:06:13]Similarly, if we constantly rely on AI to perform cognitive tasks for us, tasks that require thinking, analysis, and creativity, our own mental capabilities may not receive the necessary exercise.

[00:06:31]The convenience offered by AI makes this offloading very tempting.

[00:06:37]Tasks become faster and easier.

[00:06:40]However, this ease comes with a potential cost. Over-dependence on AI for thinking can lead to a form of mental laziness.

[00:06:51]Our brains might become less practiced at independent problem-solving, critical analysis, and original thought.

[00:07:00]The chapter emphasizes that just like physical muscles, our thinking skills need to be used regularly to stay vibrant and capable in a world that still requires human judgment and intelligence.

[00:07:15]Four, letting a computer decide.

[00:07:19]In the last chapter, we learned how using tools can make us use our brains less, a process called cognitive offloading.

[00:07:29]But this does not only happen with simple tasks. It can happen even when important decisions are made. Let's look at a real story.

[00:07:40]In the year 2023, the police in the city of Detroit were looking for someone.

[00:07:47]They used a computer program for facial recognition.

[00:07:51]This program compares faces from videos or pictures to a database of photos.

[00:07:59]The computer found a match. The police trusted the computer's result completely. Based only on this match, they went to arrest the person the computer found. But the computer was wrong.

[00:08:14]The person was a woman named Porcha Woodruff. She was 8 months pregnant at the time and had not done the crime the police were investigating.

[00:08:25]The match was from an old photo from years before for a small offense.

[00:08:31]Still, because the computer said it was her, the police arrested her.

[00:08:36]This was a wrongful arrest. It shows the big problem with trusting a computer too much, especially when the computer can make mistakes.

[00:08:47]It highlights the danger of letting AI decide without human checking or doubt.

[00:08:55]This idea of letting computers decide affects our daily lives, too.

[00:09:00]Think about platforms like YouTube, Instagram, or others you use online.

[00:09:06]They use algorithms.

[00:09:08]These are computer programs that choose what videos, photos, or posts you see.

[00:09:15]They learn what you like and show you more of that. It makes using these apps easy and comfortable. You don't have to search hard for new things.

[00:09:25]The computer brings them to you. But are you choosing what you want to see and do? Or is the computer's algorithm making the choices for you? This is called algorithmic complacency.

[00:09:39]We become comfortable letting the computer decide what information we get or how we spend our time online.

[00:09:47]We stop making these decisions ourselves.

[00:09:52]We give up our choice to the algorithm because it is easier. This is another way we might use our own thinking less.

[00:10:00]Letting a computer decide might seem helpful, but it means our brain muscles for choosing and thinking are not used as much. Five, when AI is wrong.

[00:10:12]In earlier parts, we saw how using tools too much can change how we use our brains.

[00:10:20]We also saw how trusting AI completely can lead to problems like wrong arrests.

[00:10:26]But there is another big issue we must understand. Sometimes the AI is simply wrong.

[00:10:33]Think about some answers AI has given people.

[00:10:38]A well-known AI search tool once suggested people should eat one rock each day for health.

[00:10:45]This is a very bad and dangerous idea.

[00:10:49]It also said that snakes are mammals, which is completely false.

[00:10:55]These are just two real examples from recent times. There have been many more instances where AI gave incorrect information.

[00:11:05]When AI gives answers that are not true or seems to make things up that do not exist in reality, people call this AI hallucinations.

[00:11:17]AI language models are built to predict the next word based on patterns in data.

[00:11:24]They do not truly understand facts or the world.

[00:11:28]This means they can easily produce information that is confidently stated but completely wrong.

[00:11:35]Many people are starting to trust what AI says very much, sometimes as much as human experts.

[00:11:44]But investigations have shown that many AI-generated summaries of news and information contain significant errors.

[00:11:54]Even using AI for simple tasks like rewriting text to sound better can accidentally change the original meaning and a person might not notice.

[00:12:06]A serious problem for the future of AI and online information is called model collapse.

[00:12:15]AI learns by processing massive amounts of text and data from the internet.

[00:12:21]However, increasingly, content found on the internet is being created by AI itself.

[00:12:29]Think about what happens when AI learns from content that was also made by AI, especially if that first AI was not perfect or included hallucinations.

[00:12:42]Studies have looked at this.

[00:12:45]They found that when AI learns from text that another AI generated, the quality of the new AI's outputs gets worse rapidly.

[00:12:55]After training on AI-made content just a few times, the information it produces can become strange, lose facts, or even turn into nonsense.

[00:13:08]This occurs because the AI is essentially learning from flawed data created by other AIs.

[00:13:17]It's like the AI is learning from its own kind's mistakes, and the quality of information degrades over time.

[00:13:25]This model collapse is a serious danger.

[00:13:30]Estimates suggest that a large portion of internet content today might already be generated by AI.

[00:13:38]If this AI-made content is flawed or contains errors, it means the source material for training future AI models is becoming poorer.

[00:13:50]This could lead to a rapid decline in the accuracy and reliability of the information that AI tools provide to people.

[00:13:59]This makes it crucial to remember that current AI models are not perfect sources of truth.

[00:14:06]They can be wrong. They can hallucinate.

[00:14:10]And the quality of their information might even get worse due to model collapse.

[00:14:16]We cannot simply trust AI outputs without question.

[00:14:21]We must use our own thinking and check the information it gives us.

[00:14:26]Six. An internet of lies. AI learns by reading huge amounts of text and data from the internet.

[00:14:35]But a serious problem happens when the information AI learns from is not accurate or is low-quality.

[00:14:43]This can cause something known as model collapse.

[00:14:48]This means the AI's performance and the quality of its output get worse over time.

[00:14:54]It happens because the AI starts learning from data that is not reliable.

[00:15:00]And sometimes this bad data was even created by other AI.

[00:15:06]Think of it like making many copies of a copy of a document. With each new copy, the text becomes less clear and accurate.

[00:15:17]After many steps, the words might be hard to read or understand. The AI's information becomes less and less true.

[00:15:27]This situation raises concerns about what is called the dead internet theory.

[00:15:33]This idea suggests that a big and increasing part of the internet's content is not made by people anymore.

[00:15:41]Instead, it is created by AI programs.

[00:15:47]One study has suggested that perhaps over 60% of the content found on the internet today might already be AI-generated.

[00:15:59]If AI creates a lot of content and future AI systems then learn from this growing pool of content, much of which might be flawed or low quality, the internet could become filled with text written by machines that is often incorrect or misleading.

[00:16:19]Geoffrey Hinton, a highly respected figure known as the Godfather of AI, has pointed out a fundamental limitation.

[00:16:30]He stated that the current AI language models do not truly grasp the difference between truth and lies.

[00:16:38]These programs are designed to predict the most likely next word in a sequence based on statistical patterns learned from their training data.

[00:16:49]They do not possess a real understanding of facts or the actual world.

[00:16:56]Because of this, they can produce text that sounds very convincing and fluent but contains information that is factually wrong.

[00:17:06]If a large amount of the internet is being produced by AI that cannot tell truth from lies, and this AI is becoming less accurate over time because of model collapse. It becomes very hard to find reliable information online.

[00:17:26]People might easily start to trust and accept things that are not true simply because they read them on the internet.

[00:17:34]This highlights a significant danger of depending too much on machines that lack the ability to distinguish between fact and fiction.

[00:17:45]It suggests a future where the internet could become a place largely filled with content that is inaccurate or false.

[00:17:54]Seven.

[00:17:56]A tool, not a boss. Today, people use AI a lot in their work. Surveys show that young workers use these tools very often. More than half of them say it helps with hard parts of office life, like writing emails.

[00:18:13]AI can help businesses grow and help people work together better. It can make many tasks faster. This shows that AI can really increase how much we can do.

[00:18:26]But we must be careful about using AI too much.

[00:18:31]Using it for everything, especially to avoid thinking, is not good. If people rely too much on AI, their ability to think for themselves can become weaker.

[00:18:43]It's like a muscle that doesn't get used. Think about tools from the past.

[00:18:49]In the year 1979, a new program called VisiCalc was made.

[00:18:57]It was the first spreadsheet program for computers.

[00:19:01]Before VisiCalc, if one number in a big list changed, you had to do all the math again by hand. It took a very long time.

[00:19:10]VisiCalc changed this. It made working with numbers much faster and easier.

[00:19:17]People started buying computers because of it. Accountants saw it and were amazed because it helped them so much.

[00:19:26]But VisiCalc did not make accountants stop thinking.

[00:19:30]It was a tool that helped them work more efficiently. They still needed to understand the numbers and the math. The program was a help, not a replacement for their knowledge. AI should be like VisiCalc. It is a tool that can help us work faster and smarter. It can help us with many tasks, but it should not do all the thinking for us.

[00:19:55]We must use AI as a helper, a companion.

[00:19:59]It is not meant to be our boss. Remember that AI models are just using patterns from data. They might give answers even if they don't really know the truth. So, we need to check the information and use our own brains. The key is to use AI to improve our work, not to replace our ability to understand and think critically.

[00:20:23]Eight.

[00:20:25]Thinking for ourselves.

[00:20:27]In the past, people worried about new tools, too. Think about the year 1988.

[00:20:35]Teachers had a problem with calculators in schools.

[00:20:39]They said young students should learn math facts and how to solve problems first.

[00:20:45]Children needed to understand the basic ideas of math clearly.

[00:20:51]After that, they could use a calculator to help them work faster.

[00:20:56]The tool should be an assistant, not something that does all the work for them.

[00:21:02]This old idea is very important for us now with AI.

[00:21:07]AI is a very powerful tool for many tasks. It can help us in our jobs and studies, like other helpful tools from history that made work easier.

[00:21:20]But we must remember the lessons from the past about using tools wisely.

[00:21:25]We need to use AI as a tool. It should not be our boss. We should not let it do all our thinking for us.

[00:21:34]Our ability to think is important, like a muscle we need to exercise often.

[00:21:40]If we always let AI solve problems, our own skill might get weaker over time.

[00:21:48]We need to keep our ability to think deeply and solve complex problems by ourselves.

[00:21:55]Humans are special and unique.

[00:21:58]We have real-life experiences that shape how we understand things.

[00:22:03]We can understand the world and other people in many different ways that AI cannot. We have critical thought.

[00:22:12]This means we can look at information from AI or other places and decide if it is true or good.

[00:22:20]AI can sometimes give wrong information as we know. So, using our own judgment is very important.

[00:22:29]A famous thinker named René Descartes said something very important a long time ago. He said, "We think, therefore we are."

[00:22:40]This simple idea tells us that thinking for ourselves is a key part of what it means to be a human being.

[00:22:48]We should value our own minds and our unique human abilities.

[00:22:53]We should use AI to help us, but not let it take over our capacity to think.

[00:22:59]It is important to keep thinking for ourselves. It is the human way.

[00:23:05]What do you think about this idea?

[00:23:08]How can we use AI in a smart way for the future and still keep our own minds strong? Tell us your thoughts.

[00:23:18]Hey there. Just a second before you go.

[00:23:20]If you learn something new today, please give this video a like. Write your thoughts in the comments and subscribe to the channel. It only takes a second and it really, really helps to create even better English videos for you. Your support means the world. Thank you.