Intelligence, Large Language Models, and the Left Hemisphere

Making sense of our new AI friends

These new large language models, or LLMs, are ridiculously smart. I find myself spending a lot of time debating ideas with AI. There are some very obvious problems, like the ingratiation, the training on largely exoteric knowledge, and the political guardrails. But even with all that nonsense, I can often get a better answer from AI than I can from most graduate students and even faculty outside their specialty area.

But AI is also goofy. It hallucinates, it creates false narratives, and it does so with psychopathic levels of self-confidence. LLMs can also pivot on a dime. For example, I can work through a question from one theoretical lens and then say, “Hey, let’s try it from a psychodynamic perspective,” and in the blink of an eye it’s done. So it’s intelligence, but it’s a strange, unreliable, and flighty kind of intelligence.

One analogy that popped into my head was the left hemisphere. Talking to LLMs is like talking to a disembodied left hemisphere, or maybe a rogue left hemisphere that jumped onto silicon and left the right brain behind.

There is a long history in psychology and psychiatry of split-brain research. We know from researchers like Michael Gazzaniga in psychology and Iain McGilchrist in psychiatry that the left and right hemispheres of the brain seem to be involved in somewhat different styles of processing. We don’t notice this most of the time because the two are working together. The left hemisphere is often associated with narrative construction, linear analysis, and confabulation or storytelling. The right hemisphere is more associated with holistic perception, contextual awareness, and experiential processing.

Psychologist Seymour Epstein made a somewhat similar argument in his Cognitive-Experiential Self-Theory, proposing that humans possess both a rational-analytical mode and a more intuitive-experiential mode of processing reality. As individuals, we have selves that reflect these two modes. We have a narrative self, where we tell a story about who we are, question who we are, and ponder who we are. And we have an experiential self, where we feel reality emotionally, where we resonate and vibe, where we connect, and where we experience ourselves more directly, without all the storytelling. Again, both of these are important, and ideally they work together.

Historically, at least in the West, one main concern has been the inflation of the narrative self. Carl Jung wrote about this idea a lot. But a classic cautionary tale in that lineage might be Goethe’s story of Faust.

Essentially, a runaway narrative intelligence can lead people to become detached from reality and start doing dumb things like chasing utopias. Or, if you look at it from W. B. Yeats’s poetic perspective, something like the Industrial Revolution.

Locke sank into a swoon;
The Garden died;
God took the spinning-jenny
Out of his side. — W. B. Yeats, Fragments

So my sense is that LLMs resemble a kind of left-hemisphere-style knowledge. It is sophistry. It is clever, but there is no necessary tie to reality. It is sort of like chatting with a really charming psychopath at a party. He tells you a great story, and it can be really fun to talk to him, but you probably want to check what he is saying against some reliable sources. So there can be an emptiness in purely narrative systems. I love this lyric from the Scottish singer, Mike Scott:

I’m trying to take a stance
And rise above my contradictions
But I’m just a bunch of words in pants
Most of those are fiction —The Waterboys, Long Strange Golden Road

And, interestingly, even that psychopath analogy may not be entirely off base. There is some neuropsychological work suggesting that psychopathy may involve an interhemispheric imbalance, with relatively weaker right-hemisphere systems involved in empathy, guilt, fear, and prosocial emotion, alongside relatively stronger left-hemisphere systems associated with impulsivity, stimulation-seeking, and approach behavior. (Note: I don’t want to make too much of this. I tried testing hemispheric activation in grandiose narcissism about 20 years ago but the results weren’t stable.)

So if this left hemisphere psychopathy metaphor holds, it suggests that LLMs will be metaphorically fun to talk to at a cocktail party. But maybe I want to verify what they are telling me and not fall in love with them.

When I started thinking about this topic, I thought, “Well, I’ll check with my AI and see if this makes any sense.”

The AI, of course, confirmed that I was a genius and that no one had ever had such brilliant thoughts before. Then I said, “Stop BS-ing me,” and it replied that this was actually a decent analogy, especially the association between left-hemisphere confabulation and hallucinations in large language models, but it was just an analogy.

So then I checked Google Scholar. If this this “LLM as left hemisphere model” makes any sense somebody younger and smarter than me would already be on it.

Indeed, it looks like some AI labs are starting to build systems that resemble a left-right brain architecture, pairing more linear reasoning systems with broader imaginative prediction systems. I came across one paper titled Cross from Left to Right Brain: Adaptive Text Dreamer for Vision-and-Language Navigation, where the researchers explicitly describe a model inspired by human-like left-right brain architecture. In their framework, the “left brain” handles logical integration while the “right brain” handles imaginative prediction of future scenes.

So, if this insight has some validity, it means that when I am debating LLMs, I am exercising my more linear, narrative style of thinking. This is really fun, but it’s just sophistry, or what the Internet meme calls being a “wordcel.” My personal strategy is to balance overly left brain linguistic thinking with something more holistic and right-brained. This is not a formal practice. It basically involves walking my dogs and listening to the Dead. I meditate on existence and see what percolates up from the depths.

Links to my work: Homepage; Peterson Academy; Books on Amazon

My New Peterson Academy course: The Psychology of Wealth

My PA Intro to Psych history lecture now on YouTube

Some citations

Epstein, S. (1994). Integration of the cognitive and the psychodynamic unconscious. American Psychologist, 49(8), 709–724. https://doi.org/10.1037/0003-066X.49.8.709

Gazzaniga, M. S. (2000). Cerebral specialization and interhemispheric communication: Does the corpus callosum enable the human condition? Brain, 123(7), 1293–1326. https://doi.org/10.1093/brain/123.7.1293

Hecht, D. (2011). An inter-hemispheric imbalance in the psychopath’s brain. Personality and Individual Differences, 51(1), 3–10. https://doi.org/10.1016/j.paid.2011.02.032

McGilchrist, I. (2009). The master and his emissary: The divided brain and the making of the Western world. Yale University Press.

Yeats, W. B. (1928). The tower. Macmillan.

Zhang, P., Su, Y., Wu, P., An, D., Zhang, L., Wang, Z., Wang, D., Ding, Y., Zhao, B., & Li, X. (2025). Cross from left to right brain: Adaptive text dreamer for vision-and-language navigation. arXiv. https://arxiv.org/abs/2505.20897

Disclaimer:

The information in this post is provided for educational and informational purposes only. Nothing contained herein constitutes medical, psychological, psychiatric, financial, or professional advice of any kind. This content is not intended to diagnose, treat, cure, or prevent any condition, nor should it be relied upon as a substitute for consultation with qualified medical, mental health, financial, or other licensed professionals. Always seek the advice of an appropriately credentialed professional regarding any specific questions or concerns you may have.