badge icon

This article was automatically translated from the original Turkish version.

Blog
Blog
Avatar
AuthorKÜME VakfıNovember 29, 2025 at 6:53 AM

#16 Society and Technology Bulletin

Quote

As tools that perform our physical functions gradually take over those abilities, we begin to lose them ourselves. Although metal hand tools brought great convenience into our lives, they eroded the natural skills of our arm muscles. While calculators enable us to perform highly complex calculations, our ability to calculate without machines is atrophying. However, this substitution relationship is most strikingly evident in the artificial intelligence tools to which we are now delegating our cognitive burdens. The automation of mental activities may be accompanied by an irreversible cognitive erosion.

With the arrival of ChatGPT in 2022, large language models (LLMs) began to be widely used for both personal and professional purposes. The studies to indicates that LLM usage has led to significant increases in productivity. Yet, like many other technological tools, LLMs have limitations that must not be overlooked. What do these tools give us, and what do they take away from us?

A study conducted at MIT a study sought to answer this question by focusing on writing tasks. Participants were divided into three groups: those who relied solely on their cognitive abilities, those who used search engines for support, and those who used a language model like ChatGPT. Over three sessions, each group wrote essays using the same tool. In the fourth session, roles were reversed: the LLM group was asked to write without any tools, while the group that had relied only on cognition was instructed to use an LLM.

The research examined not only the content of the essays but also the cognitive activity during the writing process. Participants’ brain activity was measured using EEG devices; essays were evaluated by both human raters and a specialized AI system. Additionally, one-on-one interviews were conducted with participants after each session.

The findings were striking:

  • The group using LLMs showed significantly weaker brain connectivity. As external support increased, the scope of internal brain networks shrank.
  • Brain scans revealed a 47 percent reduction in neural connectivity.
  • The group relying solely on their own cognition demonstrated the highest performance, both at the neural level and in essay quality.
  • Participants who switched from using LLMs to writing without them showed a decline in cognitive engagement.
  • Those supported by LLMs showed low levels of ownership over their written texts. Of ChatGPT users, 83.3 percent could not recall or quote passages from articles they had written just minutes earlier.

This study represents a crucial early step in understanding the potential effects of LLM usage on learning. Four months of observation suggest that while AI assistance may appear to boost efficiency in the short term, it could have negative long-term consequences on learning abilities. Thus, although these tools seem highly useful for making our tasks easier, they also carry negative aspects we fail to notice at first glance—above all, the fundamental threat they pose to our cognitive capacities.

However, this study does not point to an entirely pessimistic picture. It is clearly not feasible to completely abandon all technological tools and AI models in our lives due to cognitive decline. Yet, there are reflexes we can cultivate. In particular, it is essential not to disengage from the practice of writing that accompanies the development of thought. This reflects a highly understandable trend: just as the calculator weakened individual calculation skills while accelerating the advancement of mathematics by increasing overall mathematical productivity, so too does artificial intelligence carry the potential to elevate humanity’s collective output at the cost of diminishing individual mental capacity.

If humanity does not complain about muscle atrophy after delegating physical labor to trucks, will it adopt the same relaxed attitude when it comes to mental abilities? Probably not.

The cost of the productivity and efficiency gains brought by artificial intelligence must not be the atrophy of human cognitive competence and collective cognitive decline. Precisely for this reason, we must think more deeply about how artificial intelligence should be used.

It is customary for newly elected popes of the Catholic Church to highlight certain issues that define their papacy. The newly elected American pope is demonstrating such a tendency. On his second day in office, while addressing the cardinals, he addressed the issue of artificial intelligence. In his speech to the cardinalate, he emphasized the potential dangers of AI to human dignity, labor, and justice, declaring it his papacy’s “signature issue.”

Pope XIV. Leo says he took his name from XIII. Leo, who in the Industrial Revolution era championed workers’ rights. According to the new pope, we are now at a threshold similar to that of the Industrial Revolution. Algorithms and technological platforms are harbingers of a broad social transformation.

Pope XIII. Leo is from Chicago and holds a degree in mathematics. He is more technologically adept than his predecessor, Pope Francis, yet equally cautious. Pope Leo’s emphasis on the potential dangers of AI at the center of his papacy appears to subject the close relationships established in recent years between tech giants and the Vatican to a new test. During Francis’s papacy, numerous technology entrepreneurs spent extensive time in the Vatican. Their presence indicated that the rise of technology was not entirely independent of the Vatican, or that the Vatican did not wish to remain distant from this development.

Senior executives from companies such as Google, Microsoft, and Cisco have long been flocking to Rome to win the Vatican’s favor, secure ethical endorsement, and gain public legitimacy. Yet these friendly engagements have not always progressed in the same direction. While the Vatican has called for binding international regulations on AI, Silicon Valley has not responded with comparable demands for regulation.

From Rerum Novarum to Digital Incarnation

Pope Leo’s appeal draws on a longstanding tradition. Pope XIII. Leo’s famous 1891 encyclical, Rerum Novarum, laid the foundation of Catholic social teaching in response to the social inequalities caused by the Industrial Revolution. Today’s pope is raising his voice against the digital form of the same inequality: the impact of AI on labor, the dangers of algorithmic domination, and the transfer of decision-making from humans to machines.

Pope Leo signals that the Church is prepared to use its moral authority to establish a human counterbalance against AI. He is inclined to take a more distant stance than full technological acceptance. But what exactly concerns the pope so deeply?

When considering the nature of technological advancement, one detail stands out: the fact that religious authorities have paid such close attention. This interaction suggests that technology is not as neutral a developmental journey as commonly assumed, but rather one that falls within the scope of religious vision. Therefore, recognizing these external influences and understanding the actors shaping technology is one of our fundamental responsibilities as users.

Blog Operations

Ask to Küre