Visual Literacy and AI Lesson Plans

• Composing Effective Writing Assignments
  • Examples of Effective Writing Assignments
• Scaffolding Writing to Support Student Learning
• Creating Assignments for Miami Plan Capstone Courses
• Teaching Literature Reviews
• Using Threshold Concepts to Design Assignments and Courses
• Teaching Grammar Rhetorically 
• Structuring Purposeful Group & Team Work
• Mentoring Graduate Writers

• Using ePortfolio Assignments 
• Giving Feedback to Writers
• Facilitating Meaningful Online Discussions
• Engaging Students in Effective Peer Response

• Providing Online Writing Support
• Integrating AI to Guide Learning
  • Teaching Resources and Lesson Plans

ACRL Framework for Visual Literacy Theme(s) Addressed 

• Learners perceive visuals as communicating information 
• Learners practice visual discernment and criticality 

Learning Objectives 

Students will compare and contrast two images, one created entirely by an artist and the other created, in  part, via Dall-e’s Outpainting feature in order to: 

• Explore choices made in the production of visual communications to construct meaning or influence  interpretation, especially with regard to representations of gender, ethnicity, race, and other cultural  or social identifiers. [Authority is Constructed and Contextual] [Information Creation as a Process]  [Social Justice] 
• Examine visuals slowly and deeply in order to develop and refine critical observation skills. [Research  as Inquiry] 
• Examine visuals for signs of alteration, such as cropping or use of digital filters, and consider the  intent and consequences of any changes made. [Authority is Constructed and Contextual] [Information Creation as a Process] 

Teaching Notes 

• Before the lesson, the instructor chooses a painting, or other 2-dimensional artwork, by an artist. The artwork shouldn’t be familiar through popular culture, for example, artworks like Monet’s Water Lilies  are too recognizable to be used in the lesson.  
• Artworks can be chosen from any style, time period, artist, ect. the instructor chooses. The instructor  should investigate possibilities in Dall-e Outpainting prior to class to achieve their desired results  and/or based on course content. For example, a Renaissance art history class might choose to use the  Merode Altarpiece and focus on how AI handles iconography. An art appreciation class, or a class where  students are learning formal analysis, might choose to use an artwork that has an abstract or  abstracted/stylized vocabulary to focus on formal elements.  
• This lesson has the benefit of being a compare/contrast assignment. Comparing and contrasting  artworks is one of the primary ways meaning is examined in art history and is often an area of focus on  exams and essays.  
• This assignment assumes that students already have an understanding of the formal elements and  principles of design.  
• More information on how to use Dall-e Outpainting can be found in the link in the learning objectives.
  • At this time, Dall-e still has a free account option that allows for 50 starting credits and 15  credits each month after the 50 are used. This may change.  

Activity (65 minutes) 

• Introduce students to lesson (10 mins)
  • Instructor projects predetermined artworks for students. One is a painting, or other two  dimensional artwork, created entirely by an artist. The other is the same image but instead of  being created entirely by the artist, it has been finished by Dall-e Outpainting
  • Ask: Which image do you believe was created entirely by an artist? Which image do you believe  was created entirely by Dall-e Outpainting? Use your knowledge of the formal elements and  principles of design to form your argument.
    • Briefly explain that you can upload an image into Dall-e Outpainting and it will  complete it
    • Briefly explain how AI image generators work
      • Like large language models like ChatGPT, Dall-e is trained on a large dataset (in  this case images) so that it can predict what color pixel should come next in a sequence based on the words you enter as a prompt
      • It doesn’t “know” what it’s creating, it is just predicting the next color pixel based on the dataset it’s been trained on 
  • Give students time to consider the questions and write down their initial thoughts and  responses 
  • At the end, ask students which image they think is created entirely by the artist and which is  created with the help of AI 
• Student create their arguments (25 min)
  • Put students into two groups based on their thoughts
  • Tell students they will present their arguments to the class
  • Give groups 20 minutes to formulate their arguments  
• Student debates (20 mins)
  • Each side is given 10 minutes to present their arguments 
• Wrap-Up discussion (10 mins)
  • Ask the following questions:
    • Has anyone changed their mind throughout the discussion?
    • Instructor reveals which artwork was created entirely by an artist and which was created  with the help of AI
    • Did you see any mistakes or glitches in the AI images that tipped you off or ones that  you see in retrospect? Which artwork do you think is more “successful” and why? Why?  Which artwork do you prefer? Why?
      • Note: Dall-e struggles with hands, for example
      • Note: Assure students that it can be challenging to spot the AI and that, not only  is it important they realize that, AI often makes what appear to be logical  decisions.  

Assessment recommendations 

Success for this lesson should not be based on whether students guess the correct answer on which artwork is  “real” and which is AI. Rather, success should be based on students’ reasoning behind how the formal  elements and principles of design work in each image as demonstrated through the in-class debate. AI image  generators are trained on a large dataset that translates words to pixels. Although your experience might be  different, I have found that the choices of form and design made by Dall-e’s Outpainting, while not perfect,  often make some sense.

ACRL Framework for Visual Literacy Theme(s) Addressed 

• Learners participate in a changing visual information landscape 
• Learners perceive visuals as communicating information 
• Learners practice visual discernment and criticality 

Learning Objectives 

• Students will research and discuss various AI image generators’ copyright policies in order to: ○ Prioritize ethical considerations for cultural and intellectual property when creating, sharing,  or using visuals. [Information Creation as a Process] [Information has Value] [Social Justice] ○ Consider if creation and/or use of a visual will constitute misappropriation, which dissociates  visuals from their original contexts and deprives individual generators and cultural  communities of agency and credit [Information Creation as a Process] [Information Has Value]  [Social Justice]
  • Realize that visuals in all formats are works of intellectual property. [Information has Value] 

Teaching Notes 

• The number of students in each group can vary from 3-4 depending on the size of the class
• AI image generators that can be used: Dall-e, Midjourney, DeamStudio, Fotor, etc.
  • If needed, additional AI image generators can be found through a Google search 

Activity (70 minutes) 

• Introduce activity (5 mins) 
  • Tell students class will focus on AI image generators and copyright
  • In groups of 3-4, students will research how various AI image generators handle copyright  using the AI image generator’s website, Google, and library databases. 
    • Students should answer the following questions:
      • Who owns the copyright to an image generated by the AI image generator?
      • Did you find any sources that indicate there are complications to how the AI  image generator handles copyright permissions? 
• Students research an AI image generator (20 mins) 
• Students present their findings to the class (25 mins)
  • Each group is given 5 minutes to present 
• Discussion (10 mins)
  • Ask the following:
    • How do the various AI image generators compare and contrast in how they handle  copyright?
    • Which AI image generator has the “best” copyright for users? Why?
    • Which AI image generator has the “worst” copyright for users? Why?
    • What does it mean if you didn’t find any information on copyright and your AI? and  Why might this be problematic?
    • Would you hesitate to use any of these tools based on how they handle copyright?
    • Where did you find the best information about AI image generators and copyright? Why  do you think that is? 
• Discussion of additional copyright concerns (10 mins)
  • Tell students that AI image generator copyright is currently being decided by the courts
    • Issue: Only human beings can own a copyright
    • Example: Kris Kashtanova used Midjourney to illustrate her graphic novel Zarya of the  Dawn. The US Copyright Office ruled that while she owns the copyright of the written  text and sequence of images in the novel, she does not own the copyright to the individual images made by Midjourney. 
  • Tell students that AI image generators are trained on image sets scraped from the internet,  which means that artists’ images were used in this dataset without the artists’ consent
    • Example: Getty Images is suing Stability AI for using Getty Images in the AI’s dataset  without permission. 
    • Example: Artists are also suing Midjourney for using their work without their consent.
  • Discussion Wrap-Up
    • Knowing this, would you hesitate to use any of these tools based on how they handle  copyright?
    • Would you care if images you created had been used in the dataset? Why or why not?
    • Should people be able to own the copyright to an image generated by AI? 

Assessment recommendations 

Assessment for this lesson can be based on a variety of tasks, depending on instructional goals. First,  students’ discussion. Do they have an understanding of the complexity of AI image generators and copyright?  Second, students’ presentations. Did they summarize and report their findings effectively? Third, students’  research. Did they efficiently find sources? Did they use good keywords?

ACRL Framework for Visual Literacy Theme(s) Addressed 

• Learners pursue social justice through visual practice 

Learning Objectives 

• Students will research and discuss the datasets of various AI image generators, paying particular  attention to information about policies to mitigate bias and bias in the dataset in order to: ○ Acknowledge that no platform is neutral, and that concealed factors like suggestion algorithms  and power structures within the publishing industry shape experiences with visuals. [Authority  is Constructed and Contextual] [Information Has Value] [Social Justice]
  • Anticipate the ways in which algorithms, social media, and participatory technologies obscure  or promote visuals and visual media generators, which may reflect commercial interests and  reinforce existing social dynamics. [Authority is Constructed and Contextual] [Information Has  Value] [Social Justice] 

Teaching Notes 

• The number of students in each group can vary from 3-4 depending on the size of the class
• AI image generators that can be used: Dall-e, Midjourney, DeamStudio, Fotor, etc.
  • If needed, additional AI image generators can be found through a Google search 

Activity (75 minutes) 

• Introduce activity
  • Introduce activity (10 mins)
    • Tell students class will focus on AI image generators and their datasets - specifically  focusing on bias.
    • In groups of 3-4, students will research how various AI image generators handle bias in  the AI image generator’s dataset using the AI image generator’s website, Google, and  library databases.
      • Students should answer the following questions:
        • Does the AI generator acknowledge bias in its dataset?
          • If so, what biases are present?
          • If not, can you find other sources that discuss bias in the AI image  generator’s dataset? What are the biases? 
        • What are the consequences of bias in the dataset?
          • This can either be information you found during your research or  additional consequences you consider based on your research 
        • Does the AI image generator attempt to thwart bias in its dataset? If so, how?
        • Are there other rules and/or regulations and/or restrictions in your AI  image generator’s dataset? What else did you learn about your AI image  generator’s dataset?
        • Why is it important to consider the dataset used by an AI generator?
          • This answer can be related to bias or answer the question more  generally 
• Students research an AI image generator (20 mins) 
• Students present their findings to the class (25 mins)
  • Each group is given 5 minutes to present 
• Discussion (10 mins)
  • How do the various AI image generators compare and contrast in the biases inherent to their  dataset?
  • How do the various AI image generators compare and contrast in how they handle bias in their  dataset?
  • Which AI image generator takes the most steps to limit bias in the dataset?
  • Can you think of other ways AI image generators could limit bias in their datasets? 

Assessment recommendations 

Assessment for this lesson can be based on a variety of tasks, depending on instructional goals. First,  students’ discussion. Do they understand that AI image generators have biased dataset and that this has  consequences in both the images the AI generates and how the AI functions with its dataset? Second,  students’ presentations. Did they summarize and report their findings effectively? Third, students’ research.  Did they efficiently find sources? Did they use good keywords?