Artificial intelligence has advanced rapidly in the past year — and so has speculation about how its rise might affect education. ChatGPT is the fastest-growing consumer application in history, with 100M monthly active users in January 2023, just two months after its launch. During this same time frame, the majority of K-12 teachers reported using ChatGPT, and 73% said that ChatGPT can help their students learn. AI is now smart (scores 90th percentile on the bar exam), creative (generates high-resolution images based on simple prompts), and even empathic (AI chatbots rated higher than physicians for both quality and empathy of responses among patients).
In education, Khan Academy’s Khanmigo and Quizlet’s Q-Chat show the possibilities for personalized AI tutors and teacher’s assistants. AI’s potential to disrupt assessments, instructional materials, data tools, assembly tools, and the future of learning is vast and substantial.
But important questions are lost amid the buzz about promising AI tools and ethical concerns about cheating. Who, exactly, will have access to whatever new capabilities AI brings to education? And will AI benefit students too often left behind by our current education system — particularly Black, brown, and low-income students?
Unfortunately, the history of prior technological advances in education is riddled with numerous examples of over-promising and under-delivering, with minimal impact on closing achievement gaps. Technologies from smart boards to 1-to-1 device programs to flipped classrooms didn’t fulfill their potential to change education — especially when introduced in classrooms within bureaucratically-driven institutions. The iPad fiasco in the Los Angeles Unified School District is one prominent example of this, but the mirage of technological revolutions in education is a centuries-old phenomenon dating back to Thomas Edison.
Early AI returns point to similar risks around impact and equity. Large urban districts serving higher concentrations of low-income students are prohibiting the use of AI. While some districts like New York City Public Schools are starting to support the use of AI tools by students and educators, such efforts by public school systems must also grapple with with emerging demographic divides among student AI use: a notably higher percentage of Asian and male students are using ChatGPT compared to female and rural students.
These preliminary trends are troubling for anyone focused on making our education system more responsive to the needs of students furthest from opportunity. If we don’t ensure that new technologies like AI help level the playing field for Black, brown, and low-income students, they may end up disproportionately benefiting those students who are already thriving.
The good news is that AI’s impact on education is still nascent — making this the right moment to take proactive and thoughtful measures to shape its equitable development.
At Bellwether, we’re committed to ensuring that AI works for all students. We’re grappling with fast-moving AI developments alongside school leaders, educators, and parents. We don’t have answers yet, but we suspect that our core value of tailored excellence applies: the “right” way to use AI will look different from district to district, school to school, and student to student. What works for rural communities may differ from what works for urban communities; which tools benefit students may depend on their age, academic level, or any number of other factors; what AI tools and services are available may depend on the state local funding landscape; what implementation approach works best may depend on the lived experiences of the community; and the list goes on.
As AI in the education sector evolves in the months ahead, we’ll share perspectives, tools, and resources to make sure these technological breakthroughs advance equity in education.