Artificial Intelligence is already here; we need to make access more equitable

Credit: Allison Shelley/The Verbatim Agency/EDUimages

ChatGPT is all over the news these days, but when it was first released to the public in November 2022, one of us (Linda) was completely unaware of its existence, while the other (Candice) was already exploring the ways it could be used to creatively brainstorm solutions to complex policy problems in her graduate studies.

It wasn’t until after listening to a podcast on a road trip with her two teenage sons that Linda learned about ChatGPT’s incredible ability to generate creative content, write lines of code and summarize dense literature, and that one of her sons — like 33% of 12- to 17-year-olds nationwide — had already used ChatGPT to help with school assignments.

A recent meeting of the California Collaborative on District Reform focused on the future of K-12 education further pushed our awareness of artificial intelligence in education and the efforts schools are making to prepare students for a new world. Meeting participants walked away with a better understanding of the power and limitations of AI but expressed emerging and persistent concerns around bias and equity, asking questions about how to ensure that such a powerful tool can be accessed by all students. As history tells us, new technologies often widen the gaps between the rich and the poor. More recently, research shows us that 31% of students from low-income households lacked access to technologies needed for remote learning during the Covid-19 pandemic.

So, we asked ourselves how AI can be accessed equitably — and what does that even mean?

As the academic year launches, it’s imperative that school system leaders think about how to make access to AI more equitable and empower both students and educators to navigate these tools with more critical awareness.

A haphazard approach to integrating AI into schools poses potential threats to equity. Failure to ensure access to AI in resource-limited schools potentially widens the digital divide and perpetuates unequal learning opportunities and outcomes for historically underserved students and their communities. For example, OpenAI’s GPT-4 features can only be acquired through a paid premium account, meaning the most advanced AI tools, such as analyzing images and generating graphs, might be restricted to students and communities with greater financial resources. Therefore, implementing a thoughtful, realistic approach to ensuring all students, regardless of resources, can access AI tools that are changing how we learn and work, is necessary to furthering an equity agenda.

Additionally, prioritizing equity goes beyond merely ensuring access; it requires critical awareness to integrate AI into school systems. Redefining access will require comprehensive teacher training to effectively engage with AI and integrate its many capabilities into the school and classroom. A nationwide survey revealed that 72% of K-12 teachers had not received guidance and training on integrating AI into their curriculum. But training teachers to recognize the bias inherent in the tool, learning to fact-check the results AI produces, and incorporating nuanced, human details into its output is a necessity. And more essential is ensuring that teachers in both resource-rich and resource-restricted schools have access to this training.

Understanding how AI tools are built can help shine a light on the bias and systemic issues of equity associated with AI. The 2020 documentary “Coded Bias,” for example, reveals how the quality of AI output depends entirely on the data used to train it. A recent Boston Globe story shared the experience of an Asian MIT student who asked AI to make her headshot more professional, and it gave her lighter skin and blue eyes. Demographics show that 67% of AI specialists are white and 91% are men. If AI tools learn from sources primarily produced by white males, the output generated is likely to reflect the same homogenized knowledge, insights and resulting bias.

With the rapid growth of AI technology, it is likely that AI will become increasingly integrated into schools. Students are already using AI to take notes in lectures, assist with language translation, and help solve math problems. Therefore, focused attention on redefining access is necessary to ensure that students from resource-rich schools are not the only ones with the opportunity to master AI tools that will increasingly be part of their daily lives.

We are at the beginning of a long journey of understanding and navigating the role of AI in all schools, but the conversation must begin with a thoughtful and proactive approach by system leaders to center equity and empower teachers to guide students on a pathway to more powerful learning experiences.

•••

Linda Choi is a researcher and Candice Handjojo is a research associate at the American Institutes for Research and staff members of the California Collaborative on District Reform. 

The opinions expressed in this commentary represent those of the authors. EdSource welcomes commentaries representing diverse points of view. If you would like to submit a commentary, please review our guidelines and contact us.