The proliferation of artificial intelligence (AI) in schools and workplaces has implications for workload, job security, academic integrity and skills training. But what about its impact on gender equity, asks Emily Campbell.
Dr Rebecca Marrone is an educational psychologist from the University of South Australia, specialising in the effects of AI on teaching and learning – including the gender biases implicit in AI and the need to view such technology through a feminist lens.
Inherent gender bias
Dr Marrone focused in her PhD on how to increase women’s participation in Science, Technology, Engineering and Mathematics (STEM) and make the field more equitable.
“I’ve seen similar issues with the STEM sector reflected in the field of AI,” she said.
Dr Marrone said given women only account for 10 per cent of the AI development workforce, it is unsurprising gender biases are inherent in the content produced by AI.
“When you consider AI is resourced by historical knowledge that already exists on the worldwide web, is it any wonder the technology reproduces a gender bias?” Dr Marrone said.
“The fact that AI systems can reflect historical prejudice and unconscious bias is a problem acknowledged by developers of the technology.
“However, with male voices dominating the history of AI development, there is a danger AI is destined to reproduce all the societal biases the women’s movement has fought so hard to overturn.”
Across the various touchpoints where AI is now used, Dr Marrone said there are severe ramifications for women if AI continues to be dominated by male perspectives.
“This is especially true for this generation of girls who will grow up in education systems and workplaces where AI is a partner in the process,” she said.
Dr Marrone described an experience where using Chat-GPT, ungendered questions produced responses that attributed masculine pronouns to doctors and feminine pronouns to kindergarten teachers.
“It’s a small example but gender bias is insidious and replicates all the stereotypes about women’s career options that we want to eliminate,”she said.
“Increasing the number of women working in STEM and AI can help to combat gender biases in AI by bringing diverse perspectives and experiences to the table, advocating for ethical considerations, promoting representation and providing role models for girls and young women.
“Not only do we need women’s involvement to help filter out these biases, but also to lead conversations about how the technologies can deliver better, fairer outcomes across society in areas such as health, education and wellbeing.”
PD opportunities for staff
Dr Marrone said it was important for schools and education staff to have a healthy curiosity towards AI and a basic grasp of its capabilities if they’re to teach students about it.
“AI is here to stay and being afraid or ignoring it is not going to be helpful,” she said.
“Developing educational tools to teach students to question and challenge AI is an important step in promoting critical thinking and ethical decision-making in the age of AI.
“There are ways we as teachers and students can talk and learn about AI and equip young people for the future.
“It’s about raising awareness, so we don’t blindly trust these applications and systems, so we make sure they work for us and with us.
“In the same way we have regular literacy, learning to read and write, AI literacy is learning to understand how these algorithms work – not necessarily create them ourselves, but to be critical consumers.
“Being aware of the bias or biases in the data sets is a good first step to developing a base level of people’s AI literacy.”
In terms of professional development (PD) for teachers who want to develop a foundational understanding of AI and its potential biases, plenty of options are available.
“I would suggest that teachers look for PD opportunities relevant to their subject area and teaching context, ones that provide them with hands-on experience and the opportunity to collaborate with other educators,” Dr Marrone said.
“Many online courses are available to provide teachers with an introduction to AI and its applications in education.
“For example, here at UniSA we are creating a course for teachers to upskill their AI and data literacy.
“Events like conferences, workshops and online webinars allow teachers to learn from experts in the field, network with other educators, and discover new tools and resources.
“The GRAILE – Empowering Learners for the Age of AI, AI for K-12 initiative provides teachers with a toolkit of AI-related resources they can use in classrooms.
“Industry partnerships with AI leaders are another beneficial way for teachers to access training and resources about AI and suggested the AI4ALL program which is tailored to help incorporate AI into the curriculum,” she said.
Age-appropriate lessons
There is no right or wrong age for students to start learning about AI and gender biases, with Dr Marrone stating that it will vary depending on the curriculum and students’ developmental stage.
“However, given AI is becoming increasingly prevalent, I’d encourage young students to be introduced to these concepts in an age-appropriate manner.
“In the early years, students can be introduced to basic concepts, such as how machines and computers are programmed to make decisions and how those decisions are influenced by the data the machines are trained on.
“Young children can also be asked why certain jobs or activities are often associated with specific genders and encouraged to think about why this is the case.
“In primary and middle school, students can begin to explore more complex topics, such as how AI algorithms can perpetuate gender stereotypes and the importance of ensuring the data used to train these algorithms is diverse and representative.
“In high school, students can delve deeper into the ethical implications of AI and gender biases.
“For example, upper secondary students can be taught about the potential consequences of relying on AI algorithms to make decisions that can impact people’s lives, such as in hiring or college admissions processes.
“They can also be encouraged to think about the role of bias in historical and contemporary societal issues and how AI can perpetuate or mitigate these biases.
“In general, it’s important to teach about AI and gender biases in a way that is age-appropriate, engaging, and interactive, and that encourages students to think critically about the impact of these concepts on themselves and society,” she said.
Classroom ideas
Dr Marrone outlined some examples of ways to explicitly teach students critical thinking skills using AI in class.
“Teachers can provide AI-generated text for students to read and analyse to identify any biases or inaccuracies in the text,” she said.
“Compare the AI-generated results with human-generated results and ask the students to analyse the differences or similarities and consider both approaches’ strengths and weaknesses.
“An engaging activity is to have students create their own AI algorithms to solve a problem or answer a question.
“Ask them to consider any biases that might be present in their algorithm and to evaluate the accuracy of their results.
“Evaluate the data sets used to train AI algorithms and consider any biases that might be present, then ask the students to evaluate how those biases might impact the algorithm’s accuracy.
“Another idea is for students to research and discuss the ethical implications of using AI in various contexts, focusing on the potential risks and benefits,” she said.