Psycholinguistics Reveals That AI-Generated Responses Are Getting Progressively Harder For Humans To Read And Understand

In today’s column, I examine a significant sign that generative AI and large language models (LLMs) are gradually producing written responses that are harder for humans to read and understand. Think of it this way. AI output is going beyond the reading comprehension of the everyday human. This means people won’t necessarily fully grasp what AI is telling them, nor will they necessarily act appropriately on what the AI says.

How far might this go? One viewpoint is that AI output will eventually only be comprehensible to other AIs. That seems a bit far-fetched, since AI makers will likely be astute enough to ensure the AI outputs are also usable by humans. Another possibility is that humans will up their game and match AI readability by becoming smarter and cognitively sharp. It’s an optimist’s perspective, and though a happy face turns on the weighty matter, probably isn’t in the cards. Anyway, recognizing that a gap is emerging between AI outputs and human comprehension is worth considering so we don’t get blindsided by the phenomenon.

Let’s talk about it. This analysis of AI breakthroughs is part of my ongoing Forbes column coverage of the latest in AI, including identifying and explaining key AI complexities (see the link here).

Intertwining AI And Psychology

As background, I extensively cover and analyze AI’s psychological ramifications. The rising use of AI has largely been driven by the widespread adoption of generative AI and large language models (LLMs) such as ChatGPT, GPT-5, Claude, Gemini, Copilot, Grok, etc. For a recap of my over 200 analyses and posts on AI and psychology, see the link here and the link here. I frequently speak up about these pressing matters, including in an appearance on an episode of CBS’s 60 Minutes; see the link here.

Reading AI Outputs

I’d like you to join me in a mini-experiment where you judge an AI’s responses to a relatively straightforward question. Have fun with this. It will be instructive.

The question posed to the AI is this: