Is this cognitive offloading, surrender, or stunting?

For a homework assignment students read and respond to an article about health care reform.

One student turns in two well-developed paragraphs. The writing has a tight structure, strong vocabulary, and a few thoughtful discussion-like questions at the end framed as “things I’m wondering about.”

In the generative AI age, familiar questions take new forms: Who did the work? What did they learn? What can they actually do?

Understanding three related terms provides a starting point: cognitive offloading, cognitive surrender, and cognitive stunting.

🧰 Cognitive offloading: Using a tool to support thinking

Cognitive offloading is when someone uses an external tool to handle parts of thinking. This isn’t inherently bad. A tool like a calculator can offload the need for basic computations and create time and space for higher order thinking and problem-solving.

What effective offloading looks like with AI

• Students develop their argument, pull 2 supporting quotes as evidence, use AI to suggest clearer phrasing, and then choose their final version intentionally.

• Students draft an essay, use AI to generate counterarguments to challenge their own ideas, and then decide whether or not to revise their essays.

🛑 Cognitive surrender: Letting a tool replace judgment

Steven Shaw and Gideon Nave use the term cognitive surrender for when someone adopts AI outputs with minimal to no scrutiny. In this case a tool’s fluency substitutes for the learner’s reasoning and judgment.

Offloading can say: Help me think.

Surrender says: Think for me.

What surrender looks like

• The writing is polished, but the student can’t trace their claims back to sources they reviewed.

• The student can’t explain why their argument is compelling, what assumptions it rests on, or what would change their mind.

🚫 Cognitive stunting: Bypassing skill development

Cognitive stunting is when students rely on external tools before foundational skills are built. The concern with AI, as Rebecca Winthrop argues, is that AI chatbots and companions can function as “give-me-the-answer machines” that reduce the thinking and reasoning skills needed for healthy cognitive development.

This isn’t young people getting worse at thinking. It’s them not getting enough practice to build thinking.

🧩 The tricky part

Returning to the homework assignment that opens this post, those two polished paragraphs could represent three very different realities:

• Offloading: The student did the thinking and used AI to refine expression.

• Surrender: The student relied on AI to generate the core ideas and evidence.

• Stunting: The student is outsourcing so often that core skills aren’t developing.

The takeaway? If AI can generate acceptable work on demand, then output alone is no longer evidence of learning.

🎯 So, what do we do? Aim for cognitive autonomy

Cognitive autonomy is the ability to stay in charge of one’s thinking — interpreting information, weighing evidence, and making well-reasoned decisions.

The instructional challenge is making that thinking visible. When curriculum and assessment can do this, offloading becomes safer (and more useful), surrender gets harder to hide, and stunting becomes detectable.

And if we’re serious about cognitive autonomy as a learning goal, we need more clarity on:

• The dimensions of cognitive autonomy, including evidence‑tracing, reasoning transparency, awareness of uncertainty, bias/emotion monitoring, and the ability to revise beliefs with more accurate and reliable information.

• How to measure it in real classroom tasks, so we can tell the difference between a student who can “produce” a polished response and a student who can own one.

In upcoming posts, I’ll map out these dimensions and introduce classroom-ready ways to measure cognitive autonomy. This involves extending beyond existing assessment tools and frameworks to meet the needs of a new world of thinking and learning.

Stay tuned!

🤝 Connect & Share!

If you’re finding value in these posts, help me cut through the noise to connect with others. Please share this newsletter with friends, family, fellow educators, parents, or anyone committed to better understanding how we can all future-proof our thinking.

To thinking well together!

James