The Unexpected Ways AI Might Be Shaping Our Brains
The five-year-olds could point northeast without blinking. Lera Boroditsky [https://www.youtube.com/watch?v=RKK7wGAYP6k], now a professor of cognitive science at UC San Diego, could not.
Boroditsky studies how language shapes thought. Years ago, she spent time with the Pormpuraaw, an indigenous community in Cape York, Australia. Their language doesn’t use words like “left” or “right” to navigate, but instead only cardinal directions like north, south, east, and west. They say “hello” by asking “What direction are you traveling?”
Boroditsky was lost – literally. She couldn’t orient herself. Until one day, she realized that a new “module” had opened in her mind: a bird’s-eye view of the landscape, her own location marked with a red dot.
She sheepishly mentioned this to someone in the community. The response: “Well, of course. How else would you do it?”
This story – which I first heard over a decade ago – serves as a jumping off point for our latest podcast episode – a discussion about how AI could shape our brains (whether we like it or not) and what policymakers can do about it. It has stuck with me for two reasons. The first is envy. I’ve never had a good sense of direction. The day after I got my driver’s license, I asked my mom how to get to the local library – a location I’d been driven to hundreds of times. The only navigation module I have is Google Maps, which I use to get nearly everywhere. The second reason is that this story is a prime example of how our brains can learn to do what we demand of them, whether through language or other factors.
Having GPS at the ready has likely saved me – and billions more people – collectively trillions of hours of time. Humans have long shifted our cognitive capabilities based on the tools available to us. Socrates [https://www.historyofinformation.com/detail.php?id=3439] argued that writing would weaken memory and reduce the need for face-to-face engagement. Perhaps writing did weaken memory – though writing something by hand [https://pmc.ncbi.nlm.nih.gov/articles/PMC11943480/] now might help you remember it – but it also provided us with a way to capture, preserve, and share knowledge, ideas, and information, leaving mental capacity for higher-order tasks.
But outsourcing a cognitive function to a machine or tool comes with tradeoffs – and not necessarily the ones you anticipate. Recently, the neuroscientist Vivienne Ming [https://www.linkedin.com/posts/vivienneming_were-heading-for-an-ai-fueled-dementia-activity-7448417948995842049-POfa?utm_source=share&utm_medium=member_desktop&rcm=ACoAAACO8gkBz_E5_WSi83aStnby9wdXivlzaBA] – one of my favorite thinkers on AI – highlighted research showing that building a strong internal navigation system – or spatial memory – builds the hippocampus and provides protection against dementia and Alzheimer’s. London’s cab drivers – who famously need to demonstrate their ability to navigate unaided through the entire city – have lower rates [https://www.harvardmagazine.com/2025/05/harvard-taxi-drivers-brain-health-dementia] of both afflictions than similar populations.
Ming cited this research to warn about how the use of AI could rewire our brains. LLMs allow us to “offload” far more cognitive functions than GPS or other digital tools to date – everything from research, synthesis, writing, coding, analysis, and beyond. What impact will they have on our cognitive abilities – and by extension, our health, our relationships, and our communities at large?
We don’t yet know. Some research shows [https://www.bbc.com/future/article/20260505-how-to-use-ai-without-turning-your-brain-to-mush] that over-reliance on LLMs can dull creativity, weaken critical thinking and memory, and shorten our attention spans. However, the impact may largely depend on how the technology is used [https://drphilippahardman.substack.com/p/the-cognitive-offloading-paradox] – and most of us are probably not using the technology in a way that will allow us to get the benefits. But our lack of knowledge and uncertainty in the evidence isn’t stopping many institutions, including educational institutions [https://www.nytimes.com/2026/05/05/us/usc-ai-200-million-donation.html], from going all in on the technology.
AI has rightly brought significant excitement about its possibilities to cure disease, find new treatments, and support better patient care. We’ve all just gone through a decade-long uncontrolled experiment with smartphones and social media. We’ve seen how the experiments with technology in the classroom, in particular, have gone awry. While given the hype and the potential impact, I understand why many organizations are jumping to adopt AI. But this is no time for policymakers to succumb to FOMO.
We’ll do more to address smart policy responses in the weeks and months to come – and we’d love comments and links with your ideas. How do you think AI is impacting your brain (for better or worse)? What are your strategies for dealing with it – and how do you hope your governor, mayor, and others will respond?
Spotify [https://open.spotify.com/show/3mtATdDHJppQgvGHYUdmRL] ┃Apple Podcasts [https://podcasts.apple.com/us/podcast/draft-deliberative-podcast/id1894253551]┃Pocket Casts [https://pca.st/rdw22hi3]
This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit joannamikulski.substack.com [https://joannamikulski.substack.com?utm_medium=podcast&utm_campaign=CTA_1]