Despite nearly universal AI adoption in classrooms, a massive gap exists between how essential educators believe AI skills are and how prepared they think students actually are. According to a new DataCamp report surveying over 150 teachers and 150 students, 89% of educators say AI fluency will be essential to their students' future careers, yet only 3% believe their students are currently AI-fluent. This disconnect points to a deeper problem: AI is everywhere in education, but students aren't learning how to use it strategically.

Why Are Educators So Worried About Critical Thinking?

The concerns educators express go far beyond worries about cheating. While 54% of educators listed academic dishonesty as a major concern, a much larger group, 81%, expressed worry that over-reliance on AI is eroding students' critical thinking skills. This finding suggests that educators see the real risk not as students cutting corners, but as students outsourcing the thinking process itself.

Students share some of this anxiety. Nearly half of students surveyed, 48%, said they worried that AI was doing the work for them instead of helping them learn. This self-awareness among students suggests that the problem isn't ignorance about AI's risks, but rather confusion about how to use it responsibly within their learning process.

What's Driving the Adoption Divide?

The data reveals a striking paradox: AI use is nearly universal, yet AI instruction is rare. Among students surveyed, 93% of educators and 90% of students use AI at least weekly. Yet only 29% of students said they had received substantial AI instruction from their institutions. Instead, 60% of students reported teaching themselves AI, and 86% said they would join institutional AI fluency programs if offered.

This gap between adoption and formal instruction may explain why students feel unprepared. They're using AI tools constantly, but without structured guidance on how to think critically about when, why, and how to use them effectively.

How to Build Better AI Learning in Schools

• Redesign Assessments for Transparency: Eighty-three percent of educators have already redesigned their assessments for the AI era, and the data shows why this matters. More than two-thirds of students, 68%, said they got the most value from assessments where AI use could be used openly and explained, as well as process-based work that shows drafts and reasoning. This suggests that transparency about AI use, rather than bans, helps students learn.
• Establish Clear Institutional Policies: A significant barrier to effective AI learning is confusion about the rules. Forty-three percent of students don't know if their institution has an AI policy, and 44% worry about being wrongly accused of cheating. Meanwhile, 44% of educators said their institution did not have a clear, well-communicated AI policy, with just 19% having an established, active AI governance structure. Clear policies reduce anxiety and allow students to focus on learning.
• Offer Structured AI Fluency Programs: The demand is there. Eighty-six percent of students said they would join institutional AI fluency programs if offered, yet most institutions aren't providing them. Creating formal pathways to AI literacy, rather than leaving students to teach themselves, could close the fluency gap.

Is ChatGPT Still Dominating Classrooms?

The AI tool landscape in education is shifting rapidly. ChatGPT's near-monopoly among educators has eroded significantly. In DataCamp's 2025 report, ChatGPT was used by 92% of educators, but one year later, only 72% of educators said they used ChatGPT. Meanwhile, Claude adoption doubled year-over-year, rising from 30% to 63%, and Google Gemini is close behind at 67% adoption.

This diversification matters because different AI tools have different strengths, and educators are beginning to experiment with alternatives. However, it also adds to the complexity: without clear guidance, educators and students are navigating a fragmented landscape of tools without shared best practices.

What Does the Data Say About Student Self-Assessment?

Interestingly, students have a rosier view of their own AI skills than educators do. While only 3% of educators believe students are currently AI-fluent, 27% of students self-describe as fluent. This gap between self-perception and educator assessment suggests that students may overestimate their competence, or that educators are setting a higher bar for what "fluency" means.

"When AI is thrown into a classroom without strategy, what you're left with is confusion. Our 2026 AI in Education report shows just how far reality is falling short of possibility. People aren't sure of the rules, and push boundaries, and you've got experimentation with AI without a deeper sense of how best to use it," said Jonathan Cornelissen, CEO and co-founder of DataCamp.

Jonathan Cornelissen, CEO and co-founder of DataCamp

Cornelissen's observation captures the core issue: adoption without strategy creates confusion rather than learning. The data supports this. Only 5% of educators surveyed wanted to return to a world of AI-free assessments, suggesting that educators recognize AI's value. The challenge is figuring out how to harness that value in ways that strengthen, rather than undermine, critical thinking.

The 2026 AI in Education report suggests that the path forward requires three things: clear institutional policies, structured AI literacy programs, and assessment designs that encourage students to think critically about their AI use rather than simply using AI as a shortcut. Without these elements, the gap between AI adoption and AI fluency will likely persist, leaving students unprepared for a future where AI skills are essential.

" }