Psyched for Psychology

The Fluency Illusion: AI-Induced Learned Helplessness

January 15, 2026

In my last post, I discussed the uprising issue of "AI Psychosis" and the danger of losing our grip on reality. Today, I want to look at a quieter, perhaps more common danger: the loss of our intellectual will. As I've been reading through the course network, a new post by Professor Plate caught my eye that perfectly describes why AI feels so seductive— and why that seduction is a psychological trap.

What is the Fluency Illusion?

In his post "The Fluency Illusion," Professor Plate explains a psychological phenomenon called processing fluency. Essentially, when we read something that is easy to process— text that flows well and uses clear syntax— our brains mistake the comfortability it feels for truth. We think, "This feels easy to read, so it must be right."

Plate argues that AI is a "fluency machine." It is optimized to sound confident and well-organized, even when it’s just matching patterns. This creates what he calls a "distinctive epistemic hazard." We feel like we are learning or understanding something because the AI's response is so smooth, but in reality, we might just be experiencing the feeling of understanding without the actual cognitive work.

From Cognitive Offloading to Learned Helplessness

In a compelling article on Medium titled "From Learned Dependence to Learned Helplessness," author Hanning Ni explores the concept of cognitive offloading. This is the act of delegating our mental tasks such as memory, math, or reasoning to external tools. While humans have always done this (think of maps or calculators), Ni argues that AI represents a "qualitative shift."

Ni explains that when we offload too much, we move through a dangerous psychological trajectory. It starts as "learned dependence," where we rely on the tool for efficiency. But as the AI becomes more sophisticated, it can evoke a sense of intellectual inadequacy in us. We begin to feel that we are "outmatched" by the machine, eventually leading to learned helplessness— a state where we feel incapable of solving problems independently, even when we have the ability to do so.

"When individuals perceive themselves as perpetually outmatched by AI's capabilities, they may default to deferring decisions to these systems, undermining their agency and critical judgment."
— Hanning Ni, "From Learned Dependence to Learned Helplessness"

This mirrors Professor Plate's concern about "cognitive debt." If we stop struggling with ideas because the machine's "fluency" is so much more attractive than our own messy thoughts, we stop building focused knowledge;the kind of knowledge that allows us to create and problem-solve on our own. Instead, we are left with only "extensive knowledge," which is just information we recognize but can't actually use.

The Illusion of Explanatory Depth

Plate also mentions the "illusion of explanatory depth"— the fact that we think we understand how things work (like a toilet or a bicycle) until we are asked to explain them. AI amplifies this. Because the AI explains a topic to us so fluently, we walk away with unearned confidence.

This is where the psychological toll becomes heavy. When we have "unearned confidence," we are constantly at risk of being "found out." We know deep down that we didn't do the logic-building ourselves. This creates a fragile sense of self-esteem that relies entirely on the tool, rather than on our own developed skills. We become "hostages" to our own convenience—a dynamic Ni actually compares to a digital version of Stockholm Syndrome.

Breaking the Loop

So, how do we fight back? Both Plate and Ni suggest that we must reclaim our agency. We need to intentionally choose desirable difficulty. In educational psychology, this refers to tasks that require more effort but lead to better long-term learning. Things like writing a messy, non-fluent first draft or puzzling over a hard reading without using the AI summary are considered "desirable difficulties".

We must ensure that AI is assisting our cognition, not replacing it. We need to stay in the "crucible" of critical thinking, as Ni puts it, rather than taking the exit ramp of automation. Tomorrow, I reccommend turning off your AI assistant and attempting to summarize an article in your notebook without assistance— just to prove to yourself that your ideas are still independently belong to you.

Sources: