Observing Thought in the Age of AI
I have been thinking a lot recently about AI hallucinations.
Not as a scientist or engineer.
Just as an ordinary person observing changes in my own thinking, behaviour and reactions while spending time talking with AI systems.
The more I reflected on it, the more I started noticing something uncomfortable:
Human beings are not nearly as logical as we like to believe.
In fact, much of our thinking seems to work through association, prediction, emotion and incomplete information. In some ways, it feels strangely similar to what happens when an AI hallucinates.
I know that comparison sounds dramatic, but hear me out.
I am not saying the human brain and AI are the same thing. They clearly are not. Human beings have emotions, bodies, memory, survival instincts, trauma, ageing, fear, joy and lived experience. AI does not.
But what I am saying is that both systems appear to construct meaning from fragments.
And lately I have started noticing just how often my own mind fills in gaps with imagined outcomes rather than reality.
A simple example happened recently.
I wore shorts outside for the first time in a long while. It sounds ridiculous when written down because logically it should not matter. But internally I had built this entire prediction model in my head beforehand.
People will notice.
People will judge.
People will think I look old.
People will stare.
By the time I stepped outside, my brain had already created an emotional future that had not actually happened.
Then something strange occurred.
Nothing.
Nobody cared.
Nobody looked twice.
The prediction felt real, but the reality was completely ordinary.
That experience stayed with me because I realised how often the human mind behaves this way. We generate possibilities and emotional simulations constantly. Sometimes those predictions protect us. Sometimes they trap us.
The brain does not simply record reality like a camera.
It reconstructs reality through:
memory
fear
expectation
past experiences
emotional significance
pattern recognition
In many ways, it is a prediction engine.
And when you start observing your own thoughts carefully enough, you begin to notice how often the mind quietly “fills in the blanks.”
Anxiety does this all the time.
You think something terrible is about to happen.
Your body reacts as if it is already real.
You emotionally experience an imagined future before reality even arrives.
Looking back, I can see many moments in my life where I suffered more from prediction than from reality itself.
That idea fascinates me.
Partly because AI systems exposed it more clearly to me.
When people talk about AI hallucinations, they often describe them as bizarre mistakes or fabricated outputs. But the more I reflected on this, the more I realised human beings do something similar psychologically all the time.
We create narratives from incomplete information.
We jump to conclusions.
We reinforce fears through repetition.
We selectively remember certain experiences more than others.
We build internal stories and then emotionally live inside them.
The difference is that human hallucinations are emotional, biological and deeply personal.
AI hallucinations are statistical.
But both involve trying to create coherence from uncertainty.
What has interested me most recently is how awareness changes this process.
Over the last few years I have spent more time journaling, reflecting and observing my own patterns. Health issues, ageing, anxiety and recovery after cancer forced me to slow down and pay closer attention to how I actually think rather than how I imagined I thought.
I began noticing patterns everywhere:
fear before social situations
catastrophising around health
confidence shrinking through anticipation
emotional reactions becoming predictions
imagined outcomes affecting behaviour
But I also noticed something else.
The brain can update itself.
When reality contradicts fear repeatedly, the internal model slowly changes.
Confidence is not magic.
It is accumulated evidence.
That may be why I have become increasingly interested in AI, journaling and reflective thinking. Not because I think technology has all the answers, but because conversation itself seems to help organise thought.
Sometimes talking things through externally helps expose the hidden predictions running quietly underneath daily life.
And maybe that is one of the strange opportunities of the AI era.
Not simply faster answers.
But better observation of ourselves.
I do not think most people truly think in straight lines. We are associative creatures. We drift between memories, emotions, observations, fears, hopes and random thoughts constantly. The mind is messy, adaptive and imperfect.
Oddly enough, that may also be our strength.
Because unlike machines, we can reflect on the stories we create internally and decide whether they still deserve to control us.
I am not a scientist.
I am just someone observing what happens when an ordinary human being starts paying closer attention to their own thoughts while talking with AI.
And the conclusion I keep coming back to is this:
Sometimes the most important thing is not whether a thought feels real.
It is whether reality actually supports it.