How Can AI in Education Go From Fascinating to Functional?
Dylan Arena, VP, Learning Science, McGraw Hill School
Development in Artificial intelligence tends to trundle along quietly for a while and then spring onto the world stage with some spectacular new capability (AI can play chess! …drive a car! …win a game show! …identify cats!) that spurs a new round of pontification about whether AI will be humanity’s downfall or deliverance. The latest spectacle comes from Generative Pre-trained Transformer (GPT) systems, which have shocked the world by producing stunning images and passable prose. In the world of education, this news was greeted with alarm bells predicting (among other things) the death of the essay assignment, but educators soon began coming up with fascinating ways to welcome generative AI systems into the classroom.
These educators recognize not only that AI isn’t going away but that learning to leverage AI will be an important skill for their learners. Companies that make educational content should be as resourceful and creative as these educators in figuring out how to work with AI. (Companies will also, of course, need to focus on building AI-powered systems with constant attention to safety, equity, transparency, oversight, and accountability.) Here are some areas where AI is already helping in classrooms.
Personalization at Scale
Over the past several decades, our student population has become more diverse along pretty much every measurable dimension: academic achievement, career goals, family background, learning needs, etc. To take only the dimension of academic achievement, NAEP results show that in 2019, roughly half of American 4th graders were reading at either below the 1st-grade level or above the 6th-grade level — an enormous variation in achievement within a single grade even before the COVID-19 pandemic. In fact, teachers often see this sort of span within a single classroom.
As learner variability has increased and teacher shortages have spread, the task of “meeting all students where they are” has become more Sisyphean. AI systems have a clear role to play in helping educators personalize learning to support all students. For example, McGraw Hill’s digital supplemental math and science program for grades 3–12, ALEKS®, uses AI to improve student learning outcomes by precisely determining what students know and what they’re ready to learn next.
Making Data Actionable
Educators have been exhorted to engage in data-driven decision-making since the days of the floppy disk, but they have rarely been given the tools they need to use data to inform instruction. And just as learner variability has increased over the past several decades, the volume of available data has swelled by several orders of magnitude. The task of making meaning from all these data is another one that is well-suited to AI. McGraw Hill Plus for PreK–12 is our effort to help educators collect, interpret, and act on data to meet their students’ needs.
Superpowers for Superheroes
Personalizing for every learner and analyzing oceans of data to inform instruction are just two of the many superhuman tasks that our society demands from our teachers. This expectation of exceptional efforts has been with us at least since 1853, when early universal-schooling proponent Horace Mann described the ideal teacher in literally angelic terms:
As a teacher of schools…how divinely does she come, her head encircled with a halo of heavenly light, her feet, sweetening the Earth on which she treads, and the celestial radiance of her benignity making vice begin its work of repentance through very envy of the beauty of virtue!
(Mann’s praise was specifically for female teachers, one of whose highest virtues in Mann’s eyes was that they could be paid much less than men.) 170 years later, we still expect teachers to work miracles. One useful way to think about deploying AI in education is to ask: How could we give teachers superpowered tools? This framing keeps the focus on the teacher, casting AI in a clearly supporting role. Innovative teachers have already begun using generative AI to tackle time-consuming tasks like producing variations of a single argument with different tones, composing routine emails to parents, and generating other snippets of text that just needs to be “good enough.” McGraw Hill’s Actively Learn, a digital supplemental curriculum platform for grades 3–12 ELA, science, and social studies, uses AI to grade short-answer responses, allowing educators to devote more time to higher-value activities like lesson planning and direct student interaction. Similarly, our K–5 intervention program Reading Mastery Transformations provides an online oral reading-fluency assessment powered by AI from SoapBox Labs that is as accurate as human scoring in a fraction of the time, freeing up precious hours for teachers to engage with their students.
With Great Power
Every new AI breakthrough reminds us of the technology’s potential to fundamentally alter our daily lives. All powerful tools deserve to be treated with respect, and educators are right to demand transparency, accountability, and thoughtfulness from providers and partners who leverage AI for learning. (Nor are educators alone in their appropriate concern: last year, the White House released a blueprint for an AI Bill of Rights intended to protect people across a range of sectors that AI might disrupt.) As an innovative provider of educational solutions, McGraw Hill uses AI to empower teachers and learners, but we also recognize the importance of ethical AI use. To help guide our own work, we have documented our commitment to ethical AI in education, which we will continue to review and refine as the field evolves.
Dylan Arena, VP, Learning Science, McGraw Hill School Group
Dylan Arena is a learning scientist with a background in cognitive science, philosophy, and statistics. He has done extensive research and development in next-generation assessment and giving meaning to learning-relevant data. Dylan developed software at Oracle, returned to Stanford for graduate school, and co-founded edtech startup Kidaptive, which was acquired by McGraw Hill in March 2021 to continue its mission of leveraging data to support learners and their teachers, parents, and other stakeholders. Dylan is or has been a youth mentor, tutor, substitute teacher, rugby/soccer/baseball coach, and advisor in the startup, nonprofit, and private-equity sectors.