On Thursday, September 28, I hosted the "Understanding AI Image Generators for Teaching and Learning" coffee hour at Morris Library, the second in the series of "Visual Literacy and Creation Coffee Hour." During this 75-minute conversation with a small but engaged group of faculty and students, we discussed questions and concerns surrounding using and teaching AI image-generating tools (such as Midjourney and Dall-E).
Participants had various levels of familiarity with AI image generators, from often using those tools for brainstorming class assignments to no experience. A student in Art and Design mentioned that the tools are good for generating basic images, especially when one is struggling with a prompt and needs inspiration for creativity. However, most of the participants had minimal experience due to three reasons. Firstly, AI image generators function differently, and the non-website functions, such as Midjourney using Discord bot commands, seem confusing and require extra effort for many users. Without users' deep knowledge and tuning, most AI image generators generate rather basic images and are often too slow to work. Paywalls have become the most crucial reason, such that some users felt reluctant to pay for a non-job related account, while others thought AI tools are unnecessary spending for their education.
To understand AI Image Generators, we discussed the name "generators" vs. "creators" and the different natures of AI vs. artists. Participants agreed that AI is not thinking and producing but only responding to commands from users, and therefore, AI generates. One might argue that the person given commends is the "creator" of "automatic" artworks. In response, one participant suggested that it is not the ability to participate in artmaking that determines creativity but the ability to refuse to participate. For example, when audiences were asked to participate in Yoko Ono's performance, Cut Piece, people could refuse to cut a small piece of Ono's cloth against the instruction. Another voice on AI vs. artists emphasized feelings and experiences. We all agreed that creativity has to involve emotions, and all creation process is an accumulation that leads and contributes to the final art. As one of the faculty who works with young learners mentioned, "AI uses everything available (online); kids see things around and develop creativity from an early age."
With the discussion on AI vs. artists came the question: Where do you think the originality of AI-generated images stands? Many agreed that making something completely original is hard. Originality is more about the dialogue of creating. For example, painting, drawing, and other forms of artmaking are responses to the world around artists; every decision made during the artmaking process contributes to the emotion conveyed by the final result. However, this did not answer questions such as: What is the limitation to this definition of originality? Is originality a factor of copyright? Where will the debate on copyright and AI art take us?
Another aspect of AI Image Generators, and all AI tools, that is always in the debate is the ethical use of AI technologies. Participants mentioned that the "moment of notice" for fake news and misinformation is getting more complicated because, as one person noticed, "it is harder and harder to tell how much stuff is pushed into your face that's not real" since questioning "if something is quoted, where is the source" often does not help with fact-checking anymore. Another questioned, "How do we check the source? [There is] so much information being sent off too quickly." While going through a few fact-checking tools online, we also discussed how the perspective on what is "real" and "trustworthy" changes over time. For the longest time, people have believed that photographs capture the authenticity of events. We rely on tools such as Google Reverse Search to look for time stamps on an image nowadays, but users also noticed that not every image we search now has time stamps anymore. This unavoidable situation left the group with two questions: 1) How do we deal with ethically wrong AI productions (such as voice generators in videos)? 2) How do people, the original content creators, prove authenticity when the products are being accused of AI-generated materials?
Misinformation and fake news lead us to reflect on AI tools in academic environments, specifically, how or how not to use AI Image generators in classrooms. Some educators at the coffee hour held a positive opinion on AI tools in classrooms, saying students are allowed to use AI when AI tools are not the "crutches" of learning. On the other hand, other educators suggested that AI tools are taking away human experiences, including active learning experiences and critical thinking skills.
"When people notice the work is made by AI, everyone just loses interest." One comment at the coffee hour pointed out that "handmade" adds a layer of interest to a work, and in contrast, "AI-made" stripped the interest from many and added in the sense of "rush" or even "disrespectful" to the viewers. There were multiple times participants said that when using AI Image Generators, users put in words but not their voices in the tools. The lack of emotional engagement from the AI tool users, the muting of voices from the artists on which AI-generated works are based, and the missing materiality of the final products all lead to people feeling disrespectful when looking at a piece of artwork that was, in fact, AI-generated.
Where should users of AI Image Generators build their personal boundaries? How can we see those AI tools being used in classrooms across ages? The first step is to introduce technology without strong disapproval out of fear and teach learners to use AI tools in respectable ways. Participants generally were open to introducing AI Image Generators in their classrooms. A teacher working with young learners stated that she wants to teach students that they can use AI tools while informing them that overusing those tools could kill their creativity. As she suggested, lesson plans are essential in guiding students through AI technologies and understanding the opportunities and limitations AI tools present. Other participants also brainstormed ways to embed AI Image Generators in lesson plans, such as producing images for students to critique, creating assignments for students to experiment and comment on different functions of AI Image Generators (such as filters), brainstorming with AI tools to spark creativity, and making rooms for students to reflect on the limitation as well as benefits and boundaries when creating AI works.
The discussion concluded with broader topics and more questions. We will continue the coffee hour on AI Image Generators in Spring 2024. If any related issues interest you, please let me know. If you are interested in conversations about AI tools, join us at Tech Talk Tuesday on January 16 for further discussion on "Generative AI and Education."
0 Comments.