Although that’s an acceptable, consistent course of action, it’s not ideal. In one sense, the rationale for AI education resembles the justification for sex education: Students are always going to find ways to get access, and equipping them with information can keep them safer than sweeping reality under the rug and crossing our fingers that they’ll figure things out on their own someday. Steve, an administrator whose New Jersey school is among the few with a formal AI policy, says, “We acknowledge that AI is here, and we view ourselves as being responsible for exposing students to what it is, its capabilities, and the potential dangers and pitfalls.”
So what can an AI policy look like? In Washington, D.C., the tech-forward high school Washington Leadership Academy (WLA) created a schoolwide rubric, with scores from 0–4, to indicate the acceptable level of AI use for any given assignment. Zero means no AI whatsoever; 2 means that AI is permissible for help with drafting and revising, but students “must critically evaluate and modify any AI-generated content”; and 4 encourages students to “use AI creatively,” no holds barred. Students receive consistent guidance around the policy, and there are defined consequences for inappropriate AI use. Not every WLA student feels optimistic about AI, but they still find it “scary” that peers at other schools aren’t getting any instruction around AI.
In some cases, AI may reconfigure education more fundamentally. Adam Browning, WLA’s director of academic innovation, believes that AI will increasingly allow for more self-paced work, compelling more serious consideration of course credits based on standards rather than the number of hours spent sitting in a classroom. (“But,” he cautions, “not everything should be self-paced.”)
Going far beyond WLA, the AI-based Alpha School has forged ahead in this direction, relying entirely on devices for academic instruction. There’s clearly some demand for the approach, as families have shelled out as much as $75,000 annually to send their kids there. But given everything we know about the value of human teachers, in conjunction with the lack of serious evidence supporting education that’s delivered wholly via technology, I remain deeply skeptical of such an extreme.
Above all, pursue solutions, not shiny objects. Although WLA embraces tech, its students aren’t constantly jumping between different digital tools. Some years ago, the school encouraged individual teachers to explore all the tools they could find, which Browning says enabled them “to go into the wilderness of AI tools and come back with a plan.” Together, the WLA team narrowed the lineup to a select handful, and teachers now receive in-depth and ongoing professional development on each of their three AI platforms. Browning says it’s not necessarily about whether their tools are the absolute “best” in an abstract sense, but rather about ensuring that teachers are effectively supported in using what they have. “We believe in purposeful technology, not pervasive technology,” he affirms.
WLA also embeds everything in Canvas so that students don’t have to navigate multiple links to access content or assignments, which helps them stay confident and on task. “As long as they stay on Canvas, they will stay on Canvas,” Browning explains. “The second they have to leave, they will get lost on the web.” (And that’s at a school that requires four years of computer science!) WLA’s intentional, systemic streamlining efforts save students, teachers, and families from wasted time and logistical headaches.
And across school settings, tech provides essential supports for students with disabilities, who comprise 15 percent of the public-school population. Tools such as speech-to-text and screen readers are integral to many students’ access to the free and appropriate public education that’s both ethically just and legally mandated. And for students with chronic illnesses or mobility conditions, virtual school may be the best bet for consistent access to schooling. Such accommodations can make all the difference for millions of kids.
When it comes to specific classroom activities, administrators and educators largely agree that they should be device-based only when it provides a clear advantage. For example, Michigan teacher Amanda recently brought her 3rd graders outside to take pictures of clouds with their iPads. Next, the kids used the (free!) Photos app to label their photographed clouds by type, and then they discussed their work. Without devices, a gaggle of eight-year-olds poking their fingers toward the sky would’ve struggled to figure out which clouds their friends were looking at. In Amanda’s lesson, however, their devices helped them make real-world observations, enhanced their understanding of the science lesson, and fostered a face-to-face conversation.
Systemwide planning and ongoing professional learning are key to ensuring that all kids are getting meaningful lessons such as this one. First, goal setting should happen at the district- or school-wide level: Starting with student-outcome goals and planning “backwards” from there, what is the district or school hoping to achieve? How can learning tech help teachers reach those goals? As Browning explains, for WLA, “Tech is never the starting point. The content is the starting point.”
Once those broader goals are established and they earn buy-in, teachers need initial and follow-up training on the specific tools they’re using, including (but not limited to) a formal overview of the product. Observing other teachers effectively using a tool can be a huge help. Teacher observation rubrics and feedback cycles could include one or more components for effective deployment of learning tech. Some schools designate an individual to lead this work. As her school’s “tech chair,” Amanda does everything from coaching colleagues to working with an Apple programmer to standardize the home screens on students’ devices.
District leaders also need to become more purposeful in their purchases. Too often, educators tell me, vendors come to school boards and central offices to promote their products; genuinely wanting to help teachers and students, board members and administrators hand over the credit card. Teachers are then expected to use the product du jour, regardless of its relevance to their students’ needs. Instead, as Laura suggests, districts could conduct a survey asking teachers what they most need devices for. Based on the results, the district would provide access to a limited number of top-priority, evidence-based tools that align with defined goals.
Changes on the procurement side could also benefit students and schools. Outcomes-based contracting would align incentives by requiring goal setting up front and tying payments to student outcomes. A more ambitious—but not unrealistic—idea laid out in the Handbook of Children and Screens is to hold learning-tech companies accountable to clear industry standards, such as product alignment to developmental science and substantial evidence of benefits to students. A third party, such as Consumer Reports or a governmental body, could monitor compliance.
Deploying a more carefully curated selection of learning-tech products would save school districts millions of dollars. Such streamlining would also make it more feasible to provide support to teachers, who would then be better equipped to implement tech purposefully. Using devices only for the activities where they add value should also allow for more face-to-face time, gradually building back those rusty soft skills and the perceived value of in-person attendance. Sure, kids will still manage to horse around and cheat, as they always have, but schools can make those opportunities far fewer while also building those 21st Century Skills we keep hearing about.
