The Student Got an A+ But Did They Learn Anything?

AI can produce the answer faster than your child can think it through, and that's a problem.

A student who sat down to write an essay with the help of an AI tool, didn’t struggle to organize their own thoughts, nor felt the discomfort of not knowing what to say next.

They never had to retrieve what they learned, test it against the question, and build an argument from scratch. The output looked like learning, but the cognitive work that produces learning was skipped entirely.

As a cognitive neuroscientist, this is the pattern I find most concerning about AI in education. The risk is the quiet erosion of the thinking process itself.

What actually builds a strong thinker

Learning isn't the accumulation of correct answers.

It is the repeated exercise of specific cognitive operations: retrieving information from memory, holding multiple ideas in working memory while comparing them, inhibiting impulsive responses, evaluating evidence, and constructing arguments under constraints.

These operations are meant to be effortful. The effort is what strengthens the neural circuits involved.

Cognitive science calls this "desirable difficulty," a concept introduced by Robert Bjork in the 1990s and supported by decades of research since then.

When a task is hard enough to require genuine mental work but not so hard that the student gives up, that's where the most durable learning happens.

AI tools, by design, eliminate desirable difficulty. They provide fluent, well-structured output before the student has done the cognitive heavy lifting.

The student reads the result and feels like they understand it, but reading someone else's reasoning and generating your own are fundamentally different brain processes.

One activates recognition. The other activates retrieval, construction, and evaluation. The second is harder, slower, and far more valuable for long-term development.

The fluency illusion

A well-documented phenomenon in cognitive psychology is the "fluency illusion": when information feels easy to process, people tend to overestimate how well they understand and will remember it.

Highlighted textbook passages feel familiar on a second reading, so students assume they know the material.

The same thing happens with AI-generated content: because the output is smooth and coherent, students often believe they could have produced it themselves.

This matters because it erodes metacognition, the ability to accurately assess what you know and what you don't.

Strong metacognition is one of the best predictors of academic success. Students who can tell the difference between "I recognize this" and "I can reproduce this under pressure" study more effectively and perform better on exams.

AI-assisted shortcuts undermine this self-awareness by making everything feel equally well understood.

Using AI as a thinking partner

The conversation should not be framed as technology versus learning. AI tools are powerful, and students will use them throughout their careers.

The question is whether they develop the cognitive infrastructure to use those tools well, or whether they become dependent on them before their own thinking is strong enough to stand on its own.

An analogy helps here. A calculator doesn't make you worse at math if you already understand the underlying operations. But if you use a calculator before you've developed number sense, you end up unable to estimate, unable to spot errors, and unable to reason quantitatively without the tool in front of you. The same principle applies to AI and higher-order thinking.

The goal is sequencing: build the cognitive skills first, then layer the tools on top.

What parents and educators can do

Ask process questions rather than outcome questions. Instead of "Did you finish the essay?" try "Walk me through how you decided on your argument." This shifts the emphasis from output to thinking, and it makes it harder to coast on AI-generated material without engaging with the content.

Create AI-free zones for specific tasks. Frame it as training. Designate certain assignments, study sessions, or creative projects as spaces where the student works without AI assistance. Frame it the way an athlete frames training without equipment: you're building the raw capacity that makes the tools useful later.

Teach students to notice the difference. Help them recognize when they're thinking versus when they're consuming. A simple check: "Could you explain this to someone else without looking at the screen?" If the answer is no, the understanding is surface-level. This builds the metacognitive awareness that AI-assisted workflows tend to erode.

Normalize struggle. The discomfort of not knowing, of writing a bad first draft, of sitting with confusion before clarity arrives, these are signs that the brain is doing real work. Students need adults around them who frame difficulty as productive rather than something to be bypassed.

Where FirstSignal fits in

This is exactly the challenge that FirstSignal: Cognitive Training for the AI Era was designed to address. FirstSignal is a structured 12 live sessions program that helps students strengthen attention, reasoning, discernment, and intellectual independence in an age where polished output is easier to generate than ever.

The program doesn't reject AI. It helps students become strong enough to use it without becoming dependent on it. Across the sessions, students practice thinking clearly under pressure, regulating their internal state, and recognizing when a tool is extending their thinking versus replacing it.

The goal is to develop what matters more than the output: the quality of the thinker behind it.

Amelia Enginco-Figueroa is a Swiss-educated Cognitive Neuroscientist specializing in attention, memory, and learning. She works with students, parents, and educators to apply brain science to real-world performance challenges. Learn more at aef-cnp.com.