Why implementation is the first hurdle for AI in the classroom

Ideas & Philosophy · Published 8 October 2026 Swipe more like this →

In short: When new technology enters schools, measuring success isn't the first step. According to Kristen DiCerbo, Chief Learning Officer at Khan Academy, the real challenge lies in understanding implementation. If teachers and students aren't actually using the tools, even the most advanced AI will fail to impact learning outcomes.

In the evolving landscape of AI for education, Kristen DiCerbo suggests that district leaders focus on a specific chain of logic. Before assessing efficacy, they must first understand the extent of usage and identify the specific barriers preventing teachers from adopting new tools. Without active usage, the potential for improved learning outcomes remains entirely untapped. Leaders should prioritize determining how tools are being utilised in classrooms. This helps identify specific reasons why adoption might be stalling.

One significant obstacle identified is a lack of time. Both teachers and students often lack the necessary space to explore the capabilities of AI. DiCerbo suggests that providing even 45 minutes to an hour for educators to experiment with these tools can help them model effective usage for their students. This exploration is vital for moving beyond superficial interactions. This dedicated time allows instructors to demonstrate practical applications to their pupils. Such modelling is essential for overcoming initial hesitation.

Current research at Khan Academy is focused on two critical pillars: the 'tutoring move' and cognitive engagement. Researchers are analyzing tens of thousands of student chat transcripts to see if AI tools like Khanmigo act like human tutors—offering procedural corrections, summaries, or prompts. Simultaneously, they are measuring whether students are truly engaged, which is defined by asking meaningful questions and answering thoughtfully, rather than simply clicking through a screen. The study examines whether the AI provides procedural corrections or summaries. These specific actions are inspired by existing research on human tutors.

By focusing on these elements, researchers aim to bridge the gap between technological availability and actual educational impact. If the AI fails to mimic effective tutoring moves or if students remain disengaged, the technology is unlikely to produce meaningful learning results. The goal is to ensure the software functions as a high-quality tutor. This focus ensures that technology translates into genuine educational progress.

General information, not professional advice. Spotted an error? Flag the card or see corrections and complaints.

Intellect Plus is a visual educational discovery platform for exploring fascinating facts and ideas. Brain Food, Not Brain Rot. What is Intellect Plus?

Related

• Why the most successful school districts treat AI implementation as a marathon, not sprint
• How to integrate artificial intelligence into your classroom without losing your pedagogical soul
• Classrooms adopting AI tools call for new teacher and student skills
• Why School Principals Must Be the First to Adopt Artificial Intelligence in Classrooms
• Can artificial intelligence actually support teachers during a busy Tuesday morning?
• Could the next AI renaissance make us all superhuman?
• Can conversational AI transform classroom assessments from rigid testing into meaningful student growth?
• Can artificial intelligence bridge the global gap for the marginalized?
• Beyond automation: how humans and AI can evolve together
• Can artificial intelligence actually fix the aging power grid it threatens to overwhelm?
• Can AI video tools enhance our reality instead of just replacing it?
• What if you could search the internet with only a thought?

More in Artificial intelligence · More in Ideas & Philosophy · Latest · A–Z · All topics