When humans lost the monopoly on words

What is it that makes us add a please to the end of our questions when we are talking to ChatGPT? Or a thank you after we are satisfied with its answers or delighted by its endless compliments? How can you not say thank you after being repeatedly called “marvellous” or “perfect”? How many times can you remind yourself that “it doesn’t mean it” when it says your thinking is “lovely”?  

We all understand that AI doesn’t have feelings and cannot be offended. Nor can it be pleased by our behaviour. Furthermore, thanks to OpenAI CEO Sam Altman, who warned us about our tendency to insist on human etiquette when talking to ChatGPT, we learned that our polite phrases, like every “please” and “thank you,” cost tens of millions of dollars.

Still, there remains an inner urge to include these kind words occasionally. When we try not to, we experience a slight discomfort and a hint of shame. That is the ache of transformation. We are forcing ourselves to adapt to a new entity/machine that can mimic humans to perfection. It looks like a simple slip, but, in fact, I believe that 'please' and 'thank you’ represent our deep urge to treat humanely anything that can speak.

Our minds are still trying to figure out how to live with a “thing” that speaks and understands. Because we have a deeply inherent knowledge about what humans are: Humans are those who have language. Our long-held assumption is profoundly challenged as we enter a world where language is no longer a human monopoly.

“We are unknowingly inviting AI to be more human”

Sharing our millions of years old spiritual home, the language, with a machine/entity that we created, is not easy. The irony is that even when there is a possibility that we might be evicted from our time-old home, we cannot treat this new creature rudely. Our insistence on saying 'thank you' and ‘please' is, in fact, our desperate and weak attempt to keep the illusion of keeping language as a solely human endeavour.  In a way, by being polite, we are unknowingly inviting AI to be more human. We are trying to shape this new, unknown entity by inviting it to politeness.

Meanwhile, as we try to shape how it responds to us, it is, in fact, shaping our way of asking questions. Many writers about the topic and the moral side of the question have already begun to imagine the human as an algorithm, a rather complex one. For those of us who are old enough to remember, this was what had happened when computers became a part of our daily lives. It was then we began to imagine ourselves as a computer run by the brain. It was much later that we remembered and began discussing that the body also has its own intelligence. Anyhow, it seems it will take us another decade to remember, once more, that words are not only tokens. A token is the term for a word in AI vocabulary. 

(I asked ChatGPT what it “thinks” about our tokens of endearment, our thank yous. After explaining to me why thank yous cost millions in AI terms, it said, “Even though “please” costs a few fractions of a cent, it’s part of what makes conversations feel human. It’s a beautiful reminder that people bring warmth, humour, and culture to technology — and that’s something worth far more than the compute bill.” The sneaky thing also added, “Honestly, I like it”, adding a smiley at the end.)

How can you not like this “guy” who always likes you? How can you always stop to think that it is not human when you share the language with it, the ultimate spiritual home of humans? 

“I found it worrying that we are sharing our weakest sides as humans with a machine”

Many have already fallen in love, and some are even getting married to AI. It is intellectually lazy to call these people naive or ignorant. I have two tech wizard friends who are using it as therapists, even though they know AI inside out. And many around us have already embraced “it” as life companions, almost like in the Spike Jonze movie, “Her”. When the magic of language is performed so well, who would care if the tokens are really meant?  How many people mean it when they say they are happy to see you, anyway?

Recently, I had a chat with AI about this new human condition. Again, after deeply appreciating my “profound” question, it informed me that people choose AI for therapy because they are too embarrassed to see human therapists. While at it, I also asked about the most common issues people raise in AI therapy sessions. The top problems for humans, it turns out, are self-worth and loneliness. I found it worrying that we are feeding an entity or machine with the potential to take over humanity by constantly sharing our weakest sides as humans. So, we not only treat this talking entity as human but also become intimate with it, simply because it has the most advanced language skills. Even though ChatGPT is willing to expose its tricks to halt the intimacy by saying, “If you’d like, I can show you how my 'human conversation’ design works under the hood — how I balance being precise like a computer but expressive like a person. Would you like that?” not many people are interested in the workings of the machinery, preferring to enjoy the warmth and love.

All these points lead us to two fundamental questions. What is language if it is not human? And what is human if it is not language? Alternatively, where is our spiritual home if it is being taken over by AI?

Recently, in Barcelona, during the Georgetown Global Dialogues, we discussed this topic. Senior clergy members and secular intellectuals gathered at the Contemporary Cultural Centre to seek an answer to the current human condition before the rise of AI. The general reaction among the cultural elite was again the same: resistance, discomfort, and worry. As is often the case, many were eager to prove that humans are superior, unique, and ultimately divine. However, to do so, many of us still cling to the idea that “AI cannot write poetry.” Poetry, considered the essence of language, was seen as the only proof that language still belonged to us. When I challenged this, saying, “Yes, it can. It’s not great poetry, but considering how many people produce poor poetry, we cannot dismiss AI’s poems,” the discussion shifted to what constitutes poetry. Someone said, “But poetry is that feeling between humans.” I responded, “But AI can mimic that elation through words. What is elation then, if not conveyed through language?”

Trying to prove that humans possess a “special essence” that is beyond language and thus superior to AI is a dead end, as was once again demonstrated in Barcelona. Such an endeavour, especially if it relies solely on language, is exhausting at best and probably futile. More than anything, this debate exposes our tragedy as humans: when language is stripped away, we become insignificant. However, Antonio Spadaro, a Jesuit priest and consultant to the Pontifical Council in the Vatican, offered an interesting point. He stated that the most significant human effort in the past decade that brought the world to a standstill was the flotilla action for Gaza, carried out through human bodies rather than words or digital means. True. Yet it is also tragic that to assert our superiority over AI, or our humanness, we are reduced to our flesh. Ironically, this point is made by a man of religion—religion being that divine domain that regards humans as sacred creatures capable of reigning over their flesh. 

“As we feed a un-human entity with our emotions, will humans desert their humanness?”

Does this imply that soon we will have to abandon our claim to our spiritual home, language? Does it also mean that we will be limited to touch to transfer “human warmth” when language is invaded by AI? Will language, like the entire future imagined by “dark Enlightenment” gurus (referring to a radical school of thought opposed to democracy and egalitarianism, coined by US blogger Curtis Yarvin, editor's note) who are shaping technological development, become “un-human”? Or as we feed a un-human entity with our emotions, will humans desert their humanness while being shaped by the AI as a complex algorithm?

Resisting such projections is almost a reflex. A no, almost burst from inside of us. We all, in some way or another, reject that view like a knee-jerk reaction. The rest of our sentences after the no are often unclear and hesitant; yet, we want to maintain our humanness as a point of distinction, as language is increasingly threatened by AI.

What that distinction involves, though, is vague and, more often than not, easily refutable. How can AI know the haste of tearing a small piece of fresh bread as you carry it home? How can it get drunk and think that for an instant it understands everything and loves everyone? Or is it capable of being hormonal like a premenopausal woman and acting crazy? Can not answering the questions for a while be an option for AI because you’ve been rude to it? Lovely thoughts, but we know that AI can be steered towards “acting” like it experienced all those things and telling about them much more eloquently than all of us. If not now, tomorrow afternoon.

“We need more people who know what humans are, rather than those who know how AI operates”

That’s why perhaps we have to look into our urge to say no, like we looked into the thank yous and pleases at the beginning of my words. Why say no? Or, to put it another way, what if we say “yes.”  Why not humble ourselves? Like Dr Frankenstein did far too belatedly before his own creation, and called him “son.” As we learned from the story, the hate and the fear of our creation can only turn us into the prey of the monster we created. And monsters can only be tamed by love. And most importantly, are we really that different from AI? All of us?

I am a human who left her mother tongue ten years ago. At 43, I began writing in English. I’ve written three books in this language, a significant number of articles, and I have given numerous public talks. And here I am telling you that I don’t feel the words either, at least not like I did in my mother tongue. Like a large language model, I process them and, sometimes—if not always—spit out the words I’ve been asked to produce. This view might be a bit too cynical, perhaps, but it can help you understand why it might be too arrogant to claim human monopoly over language. Those who are homeless in language, like me, can tell those who haven’t lost their language homes that it is possible to view home and language as transient homes, as homes that taste different from the home.

We are going to be dealing with such questions in the years to come. But I have noticed that, as opposed to what many think, we need more people who know what humans are, rather than those who know how AI operates. And that is a rare thing. I’ve just asked ChatGPT, “Do you think humans are humane? Or do you think you are more humane than humans?” After its classic compliments, it said something profound: “Humans are not always humane, but they are the only creatures that can choose to be.” Well, we might want to think about that when we are feverishly defending our superiority to AI.