Translate this page

Lesson Ideas

4 Scientific Articles about AI and ChatGPT

Ready to have a lesson with your students about AI? How about focusing the lesson on how we can study it scientifically? Scientists around the world are hard at work researching the abilities and limitations of AI and ChatGPT. Here is a collection of four scientific articles suitable for middle and high school students. Each is accompanied by an introductory video and a selection of additional teaching resources.

This collection of adapted research articles provides students with approachable, empirical evidence of scientific research on the complex dimensions of artificial intelligence and chatGPT. Engage students with standards-matched adaptations, introductory video content, comprehension questions, and vocabulary to further your lesson outcomes. Each adapted article also comes with additional suggestions for activities to enhance student understanding and make the class more exciting.

History of the Field

Artificial intelligence (AI) describes the way computers and machines are able to learn from their environmental inputs in human-like ways. The roots of AI reach back to 1949, the invention of the Manchester Mark 1, which was the first computer able to store code or “memory.” In the 1950s computer scientists like Alan Turin and John McCarthy got curious about what a machine’s memory might lead to. They tested machines’ abilities to exhibit “intelligent” behavior. Computers in subsequent decades could store more code, and thus they were able to perform increasingly complex computations. By 1997 world chess champion and grand master Gary Kasparov was outmatched by an IBM computer called Deep Blue!

A machine’s intelligence is limited by the information it receives, and how well it recognizes the truth of and biases within that information. In our contemporary world of Siri and Alexa, predictive text, and customer service chatbots, people want to ensure we respect the power of intelligent technologies.

1. How can AI make online chats kinder?

Abstract: Imagine a world where everyone can talk and share their ideas. Even if they don’t agree, they listen to each other. This is what should happen in a democracy. But sometimes, when people chat online, things can get a bit mean. We wanted to see if artificial intelligence (AI) could make online talks kinder. So, we did an online experiment with pairs of people discussing a tricky topic: laws about who can own guns. One participant in each pair had an AI assistant that made suggestions. The participants could accept or ignore the suggestions. Most participants took the AI’s advice, which made people more positive and polite. Even if they disagreed, they understood and respected each other. The good thing is that the AI didn’t change the topic, it just helped people to talk to each other more kindly.

This article is suitable for middle school students. An audio version is available in English. This article includes a Lesson Idea video to engage students in learning about AI.

  • Key terms: artificial intelligence, politics, social media
  • Scientific figures: flow chart, scatter plot
  • Scientific method: experiment, field study, representative sampling, scientific modeling

2. How can we tell whether we are talking to a computer or a person?

Abstract: How can you tell if you are talking to a computer? New computer programs called language models have gotten very good at mimicking people. It can be really hard to tell if you’re talking to a person or a computer. We wanted to know how people try to recognize computer-generated text and if they could do it accurately. We learned that people unconsciously use rules of thumb to figure out whether they are talking to a computer. These are often wrong, which means that people are vulnerable to scams.

This article is suitable for middle school students. An audio version is available in English. This article includes a Research Recap blackboard video.

  • Key terms: artificial intelligence, machine learning
  • Scientific figures: scatter plot
  • Scientific method: experiment

3. How can humans and AI work together to detect deepfakes?

Abstract: Fake news is not new on the internet, and people often change images and videos for a joke. However, deepfakes aren’t only meant to make you laugh. Instead, they can spread misinformation or discredit a person or a group. As more deepfakes find their way onto the internet, we need to find the best way to detect these harmful videos. We tested whether the leading AI model or humans were better at detecting deepfakes online. We found that humans and the AI model were each good at identifying certain types of deepfakes. Maybe we could merge the abilities of both AI and humans to create the best deepfake detection model!

This article is suitable for high school students. An audio version is available in English. This article includes a Lesson Idea video to engage students in detecting deepfakes.

  • Key terms: artificial intelligence, machine learning
  • Scientific figures: bar graph
  • Scientific method: experiment, scientific modeling

4. How well can a computer think?

Abstract: What do chatbots, voice assistants, and predictive text have in common? They all use computer programs called language models. Large language models are new kinds of models that can only be built using supercomputers. They work so well that it can be hard to tell if something was written by a person or by a computer! We wanted to understand how a large language model called GPT-3 worked. But we wanted to know more than whether GPT-3 could answer questions correctly. We wanted to know how and why. We treated GPT-3 like a participant in a psychology experiment. Our results showed that GPT-3 gets a lot of questions right. But we also learned that GPT-3 gets confused very easily. And it doesn’t search for new information as well as people do. Knowing how and why large language models come up with wrong answers helps us figure out how to make even better versions in the future.

This article is suitable for middle school students. An audio version is available in English. This article includes a Research Recap blackboard video.

That’s Not All!

SJK is grateful to the Akamai Foundation whose support in 2023 allowed us to adapt these articles for your students.

Browse our full collections of adapted research articles on Artificial Intelligence, Machine Learning, Social Media, and Virtual Reality.

Title image from ThisIsEngineering from Pexels

Share this Lesson Idea

Check out this Related lesson idea

Latest Scientific Articles

We want to hear from you!

If you are a teacher and you used some of our resources in class, we want your feedback! Please fill out this Teacher Feedback survey!

Journal funding support from:

Recommended by: