This article is the first part in a series that explores the systemic and often senseless use of AI.

(“Walking the brain”, generated by ChatGPT, 2025)

Nearly three years after the launch of ChatGPT, revisiting its current and future societal impact may seem redundant. But if one number could summarize the cataclysmic change AI could bring, it would be 300 million. That is what Goldman Sachs estimates are the total number of full-time jobs that could be entirely automated by Artificial Intelligence today [1]. 

Yet AI’s repercussions on labour markets are only the tip of the iceberg: as it turns out, humans could rapidly transition into a post-cognitive era, where the very acts of thinking, learning, remembering & analyzing are outsourced to machines. In fact, those imperiled skills have been on a downward trend for over a decade already [2] [3].      

A recent piece by John Burn-Murdoch in the Financial Times outlines a worrying trend in OECD countries [2]: there is on average amongst the population a clear “decline in verbal & numerical reasoning”. That decrease can be attributed to a multitude of factors, chief amongst them is the shift from text-based to visual media and the steady erosion of reading habits across all age groups [4]. Some have even coined the term “post-literate society” to describe this phenomenon [5] [6]. These trends are eerily correlated with the advent of smartphones and social media in the early 2010s.  

So what becomes of our bruised cognitive abilities in the AI age [3]? 

“The world at our fingertips” was one of the internet’s great promises, and has arguably been beneficial to many, providing broader and more affordable access to information [7] [8]. Naturally, the power of accessing world knowledge came with its challenges, from polarization fueled by mis/disinformation [9], to the lively debate surrounding free speech [10]. But today’s internet does not function the same as yesterday’s. If using the internet required us to actively seek out information, today, algorithms, especially in social media, allow users to passively receive content that has already been curated. This leads to a reduction of users’ critical engagement [2].    

This passivity in how we consume content, let alone analyze and draw conclusions from it, will likely worsen with AI widespread use. Case in point, a study done by Pennsylvania University found that students using ChatGPT were scoring 48% better at math problems compared to other students, but when faced with a real life test with the same problems without the aid of AI, those students were scoring 17% worse than their peers [11]. This suggests that AI can boost productivity and output when available, but impedes problem solving capabilities when absent. Similar studies have pointed to the same conclusions. This trend extends beyond mathematics, affecting various academic fields, as the technology can be used for just about anything. Yet, it is worth noting that its results are variable: they can be very accurate on some subjects, but sometimes barely passable, if not plain wrong on others [12]. 

And therein lies the problem: where and how can we draw the line between proper and efficient use of AI and overreliance/misuse — if not utter dependence — on it? A 2024 survey from the Higher Education Policy Institute (light years away in terms of AI evolution), states that over half of UK university students were using the technology to help with their assignments, while 5% used it to cheat [13]. These numbers have likely increased significantly since then.

Stating that AI is “a plague upon education” [14] is a mere opinion judgment that sets us off track to answer the real question at stake: quid of the core intellectual abilities of the students & future generations? How exactly does AI influence them?

It goes without saying that universities and societies at large were taken aback by the quick raise of recent technologies, especially AI. It also goes without saying that lecturers cannot be fully aware of the extent to which students use AI for assignments and exams [15]. But that AI genie cannot be put back in its bottle. So, it falls upon universities and teachers to adapt and develop strategies for a better and more controlled use of AI, and not solely put the blame on AI technologies and their developers. 

That said, the academic world should also focus on the aforementioned issue at heart here: AI effects on students’ cognitive capacities. As should the world in general, by carefully monitoring what are the exact effects of AI on users’ brains.

The impact of AI will surely be generational, with those born into an AI-saturated world more likely to adopt it fully. The parallel with smartphones perfectly illustrates how the youth will harness the technology, for better or worse. Scientific communities continue to debate the effects of screen time on developing brains, but the observed externalities — such as loneliness [16], depression [9] and eating disorders [17] — do not bode well for a future in which AI resides in every young adult’s pocket.

Coupled technologies bring yet another problem: reasoning is an essential ability to have in contexts where access to technology is limited. But, in 2025, with 4/5G and wifi omnipresent and smartphones next to mandatory, we will rarely — if ever — find ourselves in such a context. We will soon truly carry all-encompassing cognitive extensions in our pockets – everywhere, at all times [18]. To try and determine what that means for the future people’s reasoning skills around the world, let’s extrapolate from the GPS technology. Researchers have identified a clear link between its overuse and the decline in spatial memory and navigation skills, as over 80% of the global adult population now carry a smartphone [19].   

Will our reasoning skills follow suit and decrease in an inversely proportional manner as Wifi and 4/5G coverage increases?  Should we see this as a blessing, or as a source of worry? It might after all not be all bad, as long as AI models greatly improve their output quality and rid themselves of biases.  But we still should proceed with caution and bear in mind the potential effects on our cognitive capacities.

If there was ever a bad omen for poor technology adoption — and what a future dominated by Artificial General Intelligence might look like — one need only note that Oxford University’s 2024 Word of the Year was “Brain rot”. It is defined as “the supposed deterioration of a person’s mental or intellectual state, especially viewed as the result of overconsumption of material, now particularly online content, considered to be trivial or unchallenging” [20].  Will the neologism “GPTing” (pretty-much self-explanatory, you have to admit) , muttered in campuses worldwide, be next year’s winner? One can only fear or hope so, depending on their affinities with AI… 
The solution to the cognitive challenges posed by AI? Swift, comprehensive regulations at every level of society [21], before we are left choking on Silicon Valley’s dust.

~

Sources: 

[1]: Briggs, J., & Kodnani, D. (2023). The potentially large effects of artificial intelligence on economic growth. Goldman Sachs. https://www.key4biz.it/wp-content/uploads/2023/03/Global-Economics-Analyst_-The-Potentially-Large-Effects-of-Artificial-Intelligence-on-Economic-Growth-Briggs_Kodnani.pdf

[2]: John Burn-Murdoch. Financial Times. (n.d.). Have humans passed peak brain power? https://www.ft.com/content/a8016c64-63b7-458b-a371-e0e1c54a13fc

[3]: Westfall, C. (2024, December 18). The dark side of AI: Tracking the decline of human cognitive skills. Forbes. https://www.forbes.com/sites/chriswestfall/2024/12/18/the-dark-side-of-ai-tracking-the-decline-of-human-cognitive-skills/

[4]: National Endowment for the Arts. (2024). Federal data: Reading for pleasure—All signs show a slump. Arts.gov. https://www.arts.gov/stories/blog/2024/federal-data-reading-pleasure-all-signs-show-slump 

[5]: O’Connor, S. (2024). Are we becoming a post literate society? Financial Times. https://www.ft.com/content/e2ddd496-4f07-4dc8-a47c-314354da8d46 

[6]: Bumi Virtual Indonesia. (n.d.). In a world where seconds decide futures, let your mind be both a blade and a well: Swift to strike. Medium. https://medium.com/@bumivirtualindonesia/in-a-world-where-seconds-decide-futures-let-your-mind-be-both-a-blade-and-a-well-swift-to-strike-26032d85d360

[7]: W. Hersh. (1999). “A world of knowledge at your fingertips”: the promise, reality, and future directions of on-line information retrieval. National library of medicine, Volume(Issue). https://pubmed.ncbi.nlm.nih.gov/10099643/

[8]: UNESCO. (n.d.). Digital age: Access to the internet has become vital for the free flow of information. https://www.unesco.org/en/articles/digital-age-access-internet-has-become-vital-free-flow-information

[9]: National Center for Biotechnology Information. (n.d.). Depression as a disease of modernity: explanations for increasing prevalence. PMC. https://pmc.ncbi.nlm.nih.gov/articles/PMC3330161/

[10]: Financial Times. (2024). The real history of free speech — from supreme ideal to poisonous politics. https://www.ft.com/content/d1f10dd6-501b-46fc-9c54-8b9697f0fc0f

[11]: Hechinger Report. (n.d.). Kids using ChatGPT perform worse on tests. https://hechingerreport.org/kids-chatgpt-worse-on-tests/

[12]: Chatbase. (n.d.). Is ChatGPT accurate?. Chatbase. https://www.chatbase.co/blog/is-chatgpt-accurate

[13]: Higher Education Policy Institute (HEPI). (2024, February 1). New HEPI policy note finds more than half of students have used generative AI for help on assessments, but only 5% likely to be using AI to cheat. https://www.hepi.ac.uk/2024/02/01/new-hepi-policy-note-finds-more-than-half-of-students-have-used-generative-ai-for-help-on-assessments-but-only-5-likely-to-be-using-ai-to-cheat/ 

[14]: Inside Higher Ed. (2023, February 9). ChatGPT: A plague upon education? https://www.insidehighered.com/views/2023/02/09/chatgpt-plague-upon-education-opinion

[15]: The Atlantic. (2024, October). ChatGPT vs. the university honor code. https://www.theatlantic.com/ideas/archive/2024/10/chatgpt-vs-university-honor-code/680336/

[16]: Harvard Graduate School of Education. (2024, October). What’s causing our epidemic of loneliness—and how can we fix it? https://www.gse.harvard.edu/ideas/usable-knowledge/24/10/what-causing-our-epidemic-loneliness-and-how-can-we-fix-it

[17]: American Psychological Association. (n.d.). Speaking of psychology: Eating disorders. https://www.apa.org/news/podcasts/speaking-of-psychology/eating-disorder

[18] : Vox. (n.d.). The effects of AI on the brain: How technology is changing us. https://www.vox.com/future-perfect/403100/ai-brain-effects-technology-phones

[19]: Tollefson, J. (2016). Artificial intelligence research revives its old ambitions. Nature, 531(7596), 573–574. https://doi.org/10.1038/531573a 

[20]: Oxford University Press. (2024). “Brain rot” named Oxford word of the year 2024. https://corp.oup.com/news/brain-rot-named-oxford-word-of-the-year-2024/

[21]: Fei-Fei Li. Financial Times. (2024). Now more than ever, AI needs a governance framework. https://www.ft.com/content/3861a30a-50fc-41c9-9780-b16626a0d2e8