AI in Schools: A Call for a New Kind of Proficiency

Reframing the Conversation. “From AI-Proof Assignments to AI-Proficient Learners”

v1.0 AI Proficiency Scale T Welch March 31, 2025

AI has been a “Black Swan” event—an unexpected force reshaping nearly every profession.

Yet in education, the response too often feels reactive:

• How can we make assignments AI-proof or at least AI-resistant?

• How do we detect AI use?

• How do we prevent cheating?

But what if we’re asking the wrong questions?

A New Lens on AI in Education

Instead of seeing AI as a threat to traditional teaching, what if we embraced it as a transformational tool for learning?

Whether or not we choose to engage with it, AI is already changing how students learn, communicate, create, and think.

We often say we’re preparing students for the “real world.” Well—in that real world, AI is already everywhere. To ignore it isn’t just shortsighted—it’s irresponsible.

Our role as educators must be to help students become thoughtful, ethical, and capable users of the technologies they will face throughout their lives. And that means giving them a framework—not just for AI today, but for whatever comes next.

Right now, though, too few adults are stepping up to that challenge.

What World Languages Can Teach Us

In my 40+ years working in World Languages, one of the most transformative developments was the creation of the ACTFL Proficiency Guidelines.

These guidelines shifted our focus from grade levels and course content to something more meaningful: What can learners actually do with the language?

I believe we need something similar for AI.

The question isn’t “To AI or not to AI?”—that ship has sailed.

The real question is: How can we empower students to use AI meaningfully and responsibly?

A Proficiency-Based Approach to AI

Rather than banning or limiting AI tools, what if we helped students develop increasing proficiency in using them appropriately and ethically?

Like language learning, AI proficiency is personal. It doesn’t depend on age or grade level. It depends on how effectively a learner uses the tool.

To begin this conversation, I’ve developed a version 1.0 AI Proficiency Scale, modeled on the ACTFL framework. It ranges from Novice Low to Distinguished, and is organized around six key Use Cases:

• Prompting – Asking clear and effective questions

• Productivity – Organizing, planning, and managing tasks

• Content Creation – Writing, design, media production

• Learning Support – Tutoring, research, exploration

• Reflection & Critique – Evaluating information and AI-generated content

• Ethics & Equity – Understanding bias, fairness, and responsible use

Students can develop skills in each Use Case as they progress through proficiency levels.

Real-Life Examples of AI Proficiency

Let’s look at how this framework might apply in the classroom:

• Novice Low (Content Creation + Ethics & Equity)

A high school junior uses Perplexity AI to write a last-minute book review comparing Huckleberry Finn and Daisy Miller and tries to pass it off as his own. The result shows little reflection or understanding—just AI output, copy and pasted. The teacher uses the paper to help the student understand why these are Novice Low examples, how it was obvious that it was an AI generated paper, and helps him figure out how to use Perplexity more effectively to more deeply understand the concepts in the two novels.

• Novice Low (Prompting)

A 5th grader uses ChatGPT to create a cover image for her story. When the image doesn’t match her vision, in expressing her frustration to her teacher she realizes her prompts need to be more specific. She tries again. She’s learning—but still at an early stage. Her teacher is helping facilitate her learning how to use AI ever-more effectively.

• Advanced Mid to High (Multiple Use Cases)

For an IB Bio project, a senior and her peers research unintended effects of a wastewater recycling initiative in the community. Additionally she combines additional AI tools to create a video for the city council and a podcast for local radio. She links her findings to Calculus learning goals, and gives an oral presentation in her AP Spanish class.

She shows advanced use across multiple Use Cases—but may still be at Intermediate High in Reflection & Critique, indicating room to grow.

In each case, the teacher’s role isn’t just to assess the final product. It’s to guide students toward their next steps on the proficiency scale—a nod to Vygotsky’s zone of proximal development.

Not a Replacement—An Overlay

This framework doesn’t replace subjects like math or biology. It runs alongside them, helping teachers assess how students are using AI to support learning within those subjects. Nor does the framework threaten student learning in math or biology or other classes. It helps students use available tools responsibly for a variety of purposes.

Because in the end, the difference isn’t in the tool.

It’s in how the learner uses it.

What’s Next?

In upcoming posts, I’ll share:

• Ideas for how teachers and school leaders can support AI proficiency without adding more to their plates

• Ways this framework can align with initiatives like Portrait of a Learner

For now, I’d love your feedback on a working draft of the AI Proficiency “Can Do” Statements, organized by level and Use Case.

https://docs.google.com/document/d/1Mrq06kraMyk27Vm6jx09sNg-1EaLSxoPpFB0NxdPi4E/edit?usp=sharing

Let’s stop hiding from AI.

Let’s start guiding students to use it well.

NB: As with all of my posts, this one was prepared using a variety of AI tools in an iterative process to arrive at the final post. I invite you to asses this post using the suggested proficiency scale and the use cases. Feel free to add your assessment in the comment section.