A Brave New World: The Perils of AI Image Generators for Women
Janette Hughes and Laura Morrison
Themes: Ethical challenges in using AI
Audience & Subject: General (All Grades)
Introduction
There are many benefits to using AI in education. It has the potential to transform education by promoting personalized learning experiences, fostering accessibility and inclusion, empowering educators with valuable insights, and expanding lifelong learning opportunities. Generative AI, and particularly ChatGPT, has reached critical mass adoption, with approximately 1.6 billion visits per month, faster than any other modern innovation. ChatGPT use by teachers increased by 144% in just five short months, from November 2022 to June 2023. With over 300 AI tools, a number that is rapidly growing, educators and their students have access to content creation tools (i.e. text, images, podcasts, videos), AI detection tools, graphic tools, music/audio tools, productivity tools and pedagogical tools (i.e. instant help with generating rubrics, quizzes, lesson plans). Media headlines abound about why educators should “embrace” AI in the classroom, with the caution that it needs to be used responsibly and ethically. However, there is no global, national or provincial consensus on what this actually means for teachers and their students. In this session, we look at some of the social and ethical considerations, such as data privacy and algorithmic bias, of using AI, and specifically how AI is having a negative impact on girls and women.
General Guidelines
In this session, we will explore the following topics as they relate to girls and women:
- Gender Bias in Data and Algorithms
- Underrepresentation of Women in Tech and AI Development
- Gendered Language and Voice Assistants
- Gender-Based Harassment and Abuse Online
- Limited Representation in AI Research and Development
- Privacy and Safety Concerns
Addressing these problems requires concerted efforts from policymakers, industry leaders, educators, and communities to promote diversity and inclusion in AI development, challenge biased algorithms, and create ethical guidelines for the responsible use of AI tools. By raising awareness and advocating for change, we can work towards a future where AI benefits all individuals, regardless of gender.
Activity/Activities
Activity 1: Locating Yourself on the Classroom AI Use Spectrum
Overview
Participants will use the Classroom AI Use: What’s Cheating? What’s OK? Graphic to locate their beliefs and comfort level with their students’ use of generative AI. This will be followed by a whole group discussion on the potential ethical considerations and promising practices for integrating AI responsibly and equitably in education.
Activity 2: Critical Analysis of Midjourney
Overview
Midjourney is an AI-driven program designed to allow users aged 13 and above to generate images through descriptive prompts. Similar to ChatGPT, it uses a large language model (LLM) to interpret the meaning of the words within your prompt. However, unlike ChatGPT, Midjourney does not possess a full grasp of grammar. For instance, when prompted with “fish riding a bicycle,” Midjourney may successfully generate an image matching the description. Conversely, if the prompt is “bicycle riding a fish,” Midjourney is likely to still produce an image of a fish riding a bike. This limitation arises from two factors: 1) Midjourney prioritizes the words in the prompt over grammar; 2) it relies on its training datasets. In this scenario, it might draw inspiration from the 1996 Guinness advertising campaign, which famously played on the feminist slogan, “a woman needs a man like a fish needs a bicyle.” Participants will examine how algorithmic biases that are rooted in gender, race, and heteronormativity, often generate images that are harmful to women.
Description
In this activity, participants will use a generative AI image generator to examine the gender-based biases inherent in its algorithms. They will also engage in critical analyses of some pre-generated images and identify strategies for determining whether an image is AI-generated or real. This activity will be accompanied by a discussion on how to teach these kinds of AI literacy skills to students in K-adult education.
Key Benefits
Key benefits include the development of AI literacy skills and lesson plan ideas for participants’ classes.
Possible Challenges
A possible challenge will be accessing an AI image generator that is free and accessible for all participants so they can try their own prompts. Access to these tools changes frequently, so we will ensure that we have pre-generated images to discuss as well.
Resources
- 300+ AI Tools for Teachers [spreadsheet]: An extensive database outlining tools, links, categories, and a brief description.
- AI image generators:
About the authors
name: Janette Hughes
Dr. Janette Hughes is a Canada Research Chair in Technology and Pedagogy and Professor at Ontario Tech University. She is the recipient of multiple research and teaching awards and research grants. She is widely published and author of The Digital Principal, a guide for school administrators interested in promoting technology-rich learning environments for students and teachers. Dr. Hughes is a prolific author and presenter, sharing her work nationally and internationally in prestigious scholarly and professional journals, keynote talks, and conferences. Dr. Hughes is routinely contacted by school districts, Ministry personnel and industry partners to consult on a variety of topics, including online teaching and learning, equity issues in ed tech, creating innovative learning environments, establishing Makerspaces in schools, shifting pedagogies in a digital era and how to foster the development of global (21st century) skills and competencies in K-12 and higher education.
name: Laura Morrison
Laura Morrison, Ed.D., is an Assistant Professor in the Faculty of Education at Ontario Tech U. Her work focuses on promising practices related to online education, inclusive maker pedagogies and critically digital literacies. Laura’s current grant-funded research explores the use of culturally sustaining maker pedagogies to increase engagement and learning for students typically underrepresented in STEM.
