Learning to Navigate Generative AI

How can developing our media literacy support our well-being, our relationships and our democracy?

• What is generative AI and what impact can it have on education?
• What steps can people take to verify content they consume? 
• How will generative AI change the information and media landscape?

• Students will understand what generative AI is and how it can shape the learning process.
• Students will be able to verify information encountered online. 
• Students will consider how generative AI will impact how content is created.

This lesson is designed to fit into two 50-min class periods and includes the following student materials:

• 1 PowerPoint
• 3 handouts

Artificial intelligence (AI) refers to ‘computer systems that can absorb information, process it, and respond in ways similar to humans’. ¹ The tasks AI can be trained to complete range widely, including recommending a new TV series to you based on your viewing history, driving a car, or evaluating a medical x-ray to determine whether your bone is broken. 

Generative AI is a subset of AI that can learn to create entirely new images, audio, or text using vast amounts of training data. Examples of generative AI programs that have been in the news include OpenAI’s ChatGPT, which creates text in response to questions and prompts, and DALL-E, which creates new images that correspond to a text-based prompt.

While AI-generated content may resemble art or speech created by humans, AI programs are not conscious and do not learn in the same ways humans do. These programs actually work like a sophisticated version of the auto-complete program you might have built into your text or email. They learn patterns from their training data and use that to create plausible responses to prompts. The more data they are trained on, the better they are at creating content that mimics human-generated content.

Generative AI has been used to create overviews on topics, essays, and artwork, but the information it generates is not always correct. Generative AI’s capabilities to complete some school assignments have raised questions around how schools should regulate the use of these programs and how curriculums might need to change to reflect this new reality.

Generative AI has the potential to transform the information and media landscape. Endless images, texts, videos and sounds can be created instantly, some of which are rooted in reality and others which are entirely fabricated. Understanding how to verify information is therefore important, as is understanding how generative AI works, its potential benefits and drawbacks, and the implications of its arrival for art and media industries.

• 1Sandy Ong, Great Decisions High School: Artificial Intelligence (Foreign Policy Association, March 2021), p. 2.

This lesson has two 50-minute parts. Part I introduces the topic of generative AI and encourages students to think about the impact that generative AI can have on student learning in schools. Part II helps students consider how generative AI will affect the information landscape and how they respond to these possible changes. While we encourage teachers to teach both parts, if you are pushed for time teaching this unit, then you might wish to teach Part II. If you decide to do this, be sure to give students a definition of generative AI or incorporate activity 2 from Part I into the second lesson and remove the fourth and final activity from Part II. The handouts School Approaches to Generative AI and The Impact of ChatGPT in the Classroom could be set as part of a homework task.

Depending on your students’ prior knowledge, you may need to provide them with the following definitions before or during the lesson. Consider adding the following terms to your Word Wall:

• Artificial intelligence (noun): Computer systems that can absorb information, process it, and respond in ways similar to humans. ¹
• Generative artificial intelligence (noun): A subset of AI that can learn to create entirely new images, audio, video or text using vast amounts of training data.

• 1 Ong, Great Decisions High School: Artificial Intelligence, p. 2.

For ideas on how to navigate the world of generative AI as an educator, see the PDF Navigating a World of Generative AI: Suggestions for Educators from the Harvard Graduate School of Education.

Each lesson in this unit includes a PowerPoint of student-facing slides. The PowerPoints are intended to be used alongside, and not instead of, the lesson plans because the latter include important rationales and context that teachers should familiarise themselves with before teaching each lesson. The PowerPoints include basic content and student-facing prompts from the lesson plans but are minimally designed because we expect teachers to adapt them to fit the needs of their students and class.

Explain to students that in today’s lesson they will be reflecting on generative artificial intelligence (AI) and its possible impacts on society.

First, you would like them to think about the impact of inventions. Ask students to reflect on the following statement made by Sophocles, an ancient Greek playwright, who was alive during the fifth century BCE. 

‘Nothing vast enters the life of mortals without a curse’ ¹

1. What do you think Sophocles’ statement means? 
2. How far do you agree with his statement? 
3. Can you think of inventions that have changed the world in both positive and negative ways?
4. How, if at all, does this statement relate to the Internet and social media? 

For younger students, you might wish to give them the statement ‘All new inventions can have positive and negative impacts on the world’, and adapt the questions accordingly.

Ask students to share their thoughts in pairs, before inviting some to share their ideas with the rest of the class. 

• 1This statement is used at the opening of the docudrama The Social Dilemma, which explores the impact of social media on those who use it and on society.

Next, explain to students that they will be reading some information and watching some short videos on generative artificial intelligence. As they read and watch the content, ask them to think about the potential benefits and downsides of generative AI and the impact that it can have on society.

Artificial intelligence (AI) refers to ‘computer systems that can absorb information, process it, and respond in ways similar to humans’. ¹ Generative AI is a subset of AI that can learn to create entirely new images, audio, video or text using vast amounts of training data. The public can now access generative AI programs that can create totally fabricated content instantly: ChatGPT, for example, creates text in response to questions and prompts, while DALL-E creates images. 

While AI-generated content may resemble art or speech created by humans, AI programs are not conscious. They are trained to detect patterns in data and use that ‘learnt’ information to create content in response to questions and prompts. Generative AI has the capabilities to create essays, articles, artwork and a broad range of other tasks. However, the information it generates is not always accurate. 

Generative AI programs can also be used to personalise content for different audiences. For example, an organisation could generate one version of an image for younger viewers and another for older viewers. People consuming media can also use generative AI to alter the content they see. For example, a reader could prompt a text-based generative AI program to alter the reading level of a news story.

Then, play the following two videos:

• Generative AI explained in 2 minutes (KI-Campus, 2:02)
• Show 0:00–1:07 of Generative AI Is About To Reset Everything, And, Yes It Will Change Your Life (Forbes) 

After students have read the text and watched the videos, ask them to discuss the following questions using the Think-Pair-Share strategy: 

1. What are the potential benefits of generative AI?
   • The potential drawbacks?
2. In what ways might generative AI change society? 
   • The information landscape?
3. What impact do you think generative AI could have on schools and the way people learn?
   • How might it impact people’s abilities? Their imagination?
4. What, if any, questions do you have? 

Lead a short class discussion inviting students to share their responses. You may choose to collect the questions and see if they are answered over the course of the lesson and/or use them to inform future learning and understand what information your students would like to know about generative AI.

• 1Ong, Great Decisions High School: Artificial Intelligence, p. 2.

Place students in small groups and distribute the handout School Approaches to Generative AI or project each approach one by one on the board. Ask students to read through each approach that schools could take in relation to generative AI, and to discuss the questions that follow. 

• Approach 1: Students should be banned from using generative AI. 
  
  Teachers could use the following strategies to enforce this policy:
  • Use AI detection programs and penalise students who use generative AI to complete assignments;
  • Create assignments that generative AI programs can’t complete, for example, by asking students to include personal connections in their writing or artwork;
  • Ask students to complete all writing or creative assignments under supervision during class time.
     
  1. What are the potential benefits of this approach?
  2. What are the potential downsides?
      
• Approach 2:  Students should be allowed to use generative AI for specific purposes, but not to complete entire assignments. 
  
  Teachers should specify acceptable uses (for example, writing an email or brainstorming ideas for an assignment), as well as unacceptable uses (for example, using text generated by an AI program in an essay). Teachers could use strategies listed in the first approach for preventing students from using generative AI when it is not allowed.
   
  1. What are the potential benefits of this approach?
  2. What are the potential downsides?
      
• Approach 3: Students should be allowed to use generative AI as they choose, as long as they disclose when they use it and take responsibility for the content it creates, including verifying all information.
   
  1. What are the potential benefits of this approach?
  2. What are the potential downsides?

When students have finished discussing each approach in their small groups, lead a short class discussion asking students to share their responses. You might also ask students to vote on which approach they think is preferable and/or to share alternatives.

Next, distribute the handout The Impact of ChatGPT in the Classroom. Read the article as a class and ask students to discuss the following connection questions in pairs or small groups:

1. What was the connection between ChatGPT’s solutions for the case study and the students’ solutions? 
   • How did this make the students feel? 
2. How did the students ultimately respond to the situation? 
   • How did ChatGPT stimulate creative thinking? 
3. What does this article suggest about how ChatGPT can be used in learning spaces?
   • How can it be used to solve societal problems? 
4. The writers of the article believe that AI can be ‘a tool to transform the way we think, work, and act, reflecting back to us what we will most probably say the first time around and inviting us to try again and again’. ¹
   • How far do you agree with this assessment? 
5. How, if at all, has this article impacted your views on generative AI and learning?

Then, lead a short class discussion inviting students to share their views. 

• 1Dana Karout and Houman Harouni, ‘ChatGPT Is Unoriginal—and Exactly What Humans Need’, Wired, 14 June 2023.

Finally, invite students to complete the following 3-2-1 prompt: 

1. Note down three things you have learnt about generative AI.
2. Note down two ways generative AI could be used in the classroom.
3. Note down one policy you think your class/school should adopt. 

Inform your students that in this lesson they will be reflecting on the impact that generative AI can have on the information landscape.

Ask students to reflect on the following questions in their journals:

1. In your opinion, what can you trust more: images, videos or text? Explain your answer. 
2. How do you check that the information you are consuming is accurate? 
3. How, if at all, will generative AI change the process of verifying information? 

Lead a short class discussion, inviting students to share their responses with the class.

Share the following text with students, asking them to discuss the connection questions in pairs: 

Generative AI models are able to create content instantly that convincingly mimics human-generated texts, images and videos. AI generators, however, cannot always distinguish between fact and fiction. This means they can create extremely convincing content that is actually false. People can also use these programs to intentionally create false information, such as photo-realistic images and videos, and texts, like new stories, that are not based in reality. 

The public does not have access to the full list of sources that generative AI models such as ChatGPT are trained on. These models draw on their vast amounts of training data to create texts, so their responses often cannot be traced to individual sources. When asked, ChatGPT will share citations for its responses, but these citations may not be the actual sources of the information it shared, and may even be made up.

Connection Questions: 

1. What impact might generative AI have on the content people consume, both online and offline?
2. Why does it matter what data is used to train generative AI models? 
   • What does it mean that the public does not have access to this data?
3. How might generative AI be used to spread false information?

Next, ask your students to read and evaluate two sample texts. Inform them that one was created by ChatGPT and mimics the style of an academic article, and the other is an excerpt from a real, published academic article. Distribute the handout Evaluating Texts.

Ask students to read the passages in their groups and discuss the connection questions:

1. Which text do you think comes from a genuine published article? Why? 
2. Which aspects of the content make it seem trustworthy? Assess this for each passage.
3. What steps could you take to verify these articles? 
   • What content would you research? 
   • How would you go about doing this? 

Ask students to vote on which passage they think came from a published article, inviting one or two to explain their thinking. 

Share the answer with your students: Passage 1 is from an academic article ‘Influenza vaccine coverage and predictors of vaccination among aged care workers in Sydney Australia’, which was published in the journal Vaccine. Passage 2 was generated by ChatGPT and references a fictional study.

Then, lead a short class discussion asking students to share their responses to question three. 

Next, share with your students the following three steps they can use to check information they come across online. Explain that steps are useful for detecting misinformation created by generative AI, but are also helpful for checking human-generated content.

1. Research the organisation that published the content.
2. Verify key information in the text. Copy and paste passages/data into search engines.
3. Check any citations included in the text.

Walk through the three steps together as a class for both of these passages. 

Finally, ask your students to discuss the following question using the Think, Pair, Share strategy:

How, if at all, does the information you learnt in this activity change the way you think about and/or engage content you see online or on social media?

Inform students that they will be reflecting further on how generative AI works. 

Share one of the following videos with students – one explores AI art and the other explores deepfakes:

• AI art, explained (Vox, 13:32)
• The incredible creativity of deepfakes — and the worrying future of AI (TED, 13:07)

As students watch the video, ask them to take notes on the following questions: 

1. What do you learn about how this type of generative AI works?
2. What do you find surprising, interesting and/or troubling about the content this type of generative AI model can create? 
3. What opportunities does this generative AI model provide?  
4. What would you want to know about an AI-generated image or video you saw online? 
5. Should AI-generated content be labelled? Explain your view? 
6. Generative AI models are trained on content and data that is available online and the model has not gained direct permission to use. What are the ethical implications of this? 
7. What norms should govern the creation of AI-generated images and videos? 

Then, lead a short class discussion inviting students to share their views. 

If there is time, you might invite students to share one thing they are taking away from the lesson in a Wraparound.

Explain to your students that you will co-create a sample class contract (a series of norms) that could guide the use of generative AI in a school that allows some use of generative AI. Given how rapidly school policies may change on the use of generative AI, it may be more helpful to frame creating this contract as a learning exercise rather than the selection of binding norms for your classroom. If you have a class contract or policies related to plagiarism or cheating, you may wish to re-share these documents with your students to help them generate ideas for their sample generative AI contract.

Ask your students to work together in small groups to brainstorm a list of norms that guide when and how students can use generative AI (if you have shared existing policies, ask that they align with these). 

You may wish to share a selection of the following questions with your students to help guide their thinking:

1. How could we extend our existing class norms or policies to the use of generative AI?
2. When could using generative AI disrupt the trust between teachers and students?
3. When could generative AI be a useful tool for learning?
4. When could it undermine learning a skill or concept?
5. When and how should students acknowledge their use of generative AI in their work?

Then, ask one person from each group to share their ideas for norms with the class. As students share, write down the themes you hear. Discuss and agree on a final selection of sample norms as a class.

Share the article ‘ChatGPT in the classroom — why not?’ with students. Explain that this article was written by a 16- to 17-year-old who began using AI in their studies. Ask students to note down the practical uses of AI in school work and the risks.

Inform students that they will be exploring how generative AI can impact the world of human creativity and information. Share one or more of the following resources with students, inviting them to think about the benefits and drawbacks of generative AI in these contexts. You might ask students to synthesise their learning by creating a poster, article or video. 

• Video: AI art, explained (Vox, 13:32)
• Video: What AI art means for human artists (Vox, 16:42)
• Video: The incredible creativity of deepfakes — and the worrying future of AI (TED, 13:07)
• Video: What if you could sing in your favorite musician's voice? (TED, 9:25)
• Article: ‘Generative AI and deepfakes: How artificial intelligence tools will sow disinformation’ (Thumos)
• Article: ‘How one of the world’s oldest newspapers is using AI to reinvent journalism’ (Alexandra Topping, The Guardian)
• Article: ‘Welcome to the new surreal: How AI-generated video is changing film’ (Will Douglas Heaven, MIT Technology Review)

Show students the Code.org video AI: Training Data & Bias (2:40) and ask them to take notes on what causes AI to be biased, the possible consequences of this and how the risk of bias can be reduced. Next, share content from one or more of the following articles with students and invite them to reflect on the impacts that AI can have on the criminal justice system:

• ‘Wrongfully Accused by an Algorithm’ (Kashmir Hill, New York Times) 
• ‘AI is sending people to jail—and getting it wrong’ (Karen Hao, MIT Technology Review)
• ‘Algorithms Were Supposed to Reduce Bias in Criminal Justice—Do They?’ (Molly Callahan, The Brink, Boston University)