Does AI make us less intelligent?

In the late 1970s I was a student working on my PhD thesis in linguistics at Indiana University in Bloomington, the US.  There were no such things as laptops or mobile phones, and the use of personal computers was still a rare commodity. 

Calling Tripoli via the operator was a weekly event I engaged in so that I could hear the voice of my dear parents – now departed – and they could hear mine. 

As we still had to wait for the digital IT revolution started by Bill Gates and others, we had no emails and the fastest means of communication was still the telephone and the telegram.

Of course, there was still no such thing as GPS, but we managed, with a few blunders, to travel across the vast wilderness of the US, equipped only with a map, a pencil, and hot mugs of coffee, which kept us awake while listening to country music transmitted from the radio or the cassettes.

I mention all these examples just to show the readers, especially the younger generation, how much we have technologically advanced over the last four decades and how dependent we have become on our devices.

Speaking about GPS, for example, I remember being stopped a few months ago by a young tourist in Republic Street, Valletta needing directions to get to the Upper Barrakka. I almost pointed to the place.

As she thanked me and went on her way, I noticed she was checking her mobile and walking in the wrong direction. I caught up with her to tell her that was not the way, and to my surprise she said: “Oh thank you sir, there must be something wrong with my GPS.”

My point, of course, is not that we should not trust our devices, such as GPS, because they may err, as do humans, but that we should not become too dependent upon them to the extent that we dull our own senses.

The latest, and undoubtedly, most advanced of our devices is artificial intelligence (AI). Through a labyrinth of algorithms, it mimics our human intelligence and does so with amazing speed and increasing cognitive competence.

It is built on a large network that consists of billions of parameters trained on vast amounts of texts enabling it to learn patterns in language.

Humans learn, or more accurately acquire language, by an innate capacity to construct rules, hence language is described as rule-governed behaviour, which through a finite set of rules can generate potentially an infinite number of sentences, as Noam Chomsky and other linguists have shown.

AI does not have this ability as it is pattern governed, yet it is a system that can ‘learn’ and has the ability to generate responses by predicting what text should come next based on patterns it has learned.

AI is an artificial system which is not designed to replace our own thinking- Saadun Suayeh

The human mind seems to have created a system in its own image, yet nonetheless it is an artificial system which is not designed to replace our own thinking.

However, there is an increasing tendency among many people to use it, especially through applications such as ChatGPT, as a magic tool of thinking solving virtually anything. As I set on writing this article, my grandson advised me to accelerate the process by resorting to ChatGPT. People now use it, he said, to write their emails, letters, business reports, medical reports, petitions, visa applications, term papers, and possibly even poems and love letters.

Although I am aware that many of us are becoming increasingly dependent upon their AI applications, such as ChatGPT, I am alarmed by the prospect of a ‘device’ or a ‘system’ doing so many tasks on our behalf, including perhaps writing homework for schoolchildren and columns for journalists.

Would this lead to cognitive laziness and intellectual atrophy? Or am I being pessimistic about a device that can open frontiers hitherto untrodden, or even unthinkable, for mankind?

I pose these questions not as much to provide answers for my readers as to raise their curiosity. Personally, I enjoy using my ChatGPT and even like to tease it. Being a linguist, I like to see how it handles the intricacies of language, especially in translation. I would submit to it lines from Shakespeare to translate into Arabic and lines from classical Arab poets to translate into English.

At times the translation seems semantically accurate, though poetically awkward, but can also be wrong. However, the system is improving with amazing results. In fact, we (the system and I) even engage in discussing alternatives, enhancing its learning capabilities.

However, no matter how hard we try to humanise our AI devices, they remain nothing but advanced neural networks of mathematical algorithms created by the human mind.

We should guard against overreliance on such systems and the frightening prospect of creating an Orwellian world of robots for our future generations.

The question, therefore, is not whether AI will make us less intelligent, but rather, would we choose to become so by sacrificing our intellect in an age seeking easy answers and letting the GPS charter our destination.

Saadun Suayeh is a former ambassador of Libya to Malta.