GenAI Use Linked to Lower Independent Thinking Confidence

Ever feel like you’re leaning a little too much on AI for your big ideas? You’re not alone, and a recent study suggests this reliance might be subtly affecting how confident you feel in your own brainpower. It turns out, how you *use* Generative AI could be more important than simply *using* it.

What the Study Uncovered About GenAI and Confidence

Researchers from the American Psychological Association dove deep into this very question. They published a study titled “Generative Artificial Intelligence Reliance and Executive Function Attenuation: Behavioral Evidence of Cognitive Offload in High-Use Adults.” To get a clear picture, they gathered insights from nearly 2,000 working adults across various fields like tech, finance, and healthcare. These individuals were already integrating GenAI into their daily professional tasks.

The study allowed participants to use AI tools freely for complex tasks – think planning with evolving information, prioritizing projects, or explaining strategic decisions. The key wasn’t just *if* they used AI, but *how* they engaged with its output. Did they blindly accept suggestions, or did they actively refine and question them?

The Confidence Connection: Revision is Key

Here’s where it gets interesting: participants who largely accepted AI’s suggestions with minimal changes reported a noticeable dip in their confidence regarding independent reasoning. On the flip side, those who actively modified and revised the AI’s output felt more secure in their own ability to think critically and independently.

Ethan Mollick, an associate professor at the Wharton School who studies AI’s impact on work, shared a similar sentiment with Time. He noted that if AI simply solves problems for you, you might not learn or engage your own thinking processes. However, when AI acts more like a tutor, prompting and challenging you, the outcomes can be much better.

It’s All About Your Interaction Style

The study’s lead author, Sarah Baldeo, a PhD candidate at Middlesex University, emphasized that Generative AI isn’t inherently good or bad for our cognitive abilities. It’s our interaction style that dictates whether it leads to cognitive decline or evolution. Brain activity scans showed that depending on how people choose to use these tools, we can see increases or decreases in cognitive engagement.

So, the presence of AI assistance isn’t the sole factor. It’s the patterns of oversight, authorship, and how deeply you engage with the AI’s suggestions that truly shape your perceived autonomy and confidence in your own thinking. It’s about treating AI as a collaborator you guide, rather than an oracle you simply obey.