){kind=logo}
){kind=icon}
){kind=logo}
){kind=icon}
There are so many experts warning us today about the risks of artificial intelligence that in this article, I propose to advocate for this technology, at least for a topic related to several branches of teaching, which I find especially intriguing: AI in producing creative texts.
My first impulse in addressing the issue is to highlight the hundreds of “Hollywood” film and series writers (I mean this style, whatever the country) who, in recent weeks, have protested the use of AI in the execution of tasks they believe to be only their responsibility. Arguing, among other things, that this technology standardizes and dehumanizes creative values, they seem to forget that they also have been committing this sin for decades, that is, executing scripts based on an invariably statistically proven pattern of success that ruins the originality of many stories (their own or others), turning them into cookie cutter pieces to secure the market. The fact that they now protest the system that favors machines doing the same “maquiladora” work more efficiently seems, to put it kindly, a subject for a script about one of the thousand ways human beings sell their souls to the devil.
But I leave it to them to write this script (or the chatbot that has already replaced them), and now restraining my anger (which comes from nothing but a viewer thousands of times disappointed), I will attempt to pour my energy into a more creative approach to the subject. After all, this game of siding with AI on the issue of literary creation will allow me to enjoy my fantasy to the top, which, although risky, promises unceasing pleasures.
So, my first indulgence is to compare ChatGPT with William Shakespeare. I will have time to moderate this exaggeration, but the reader will see that the comparison is not as unjustified as it first seems. For now, to make it more plausible, let’s imagine that we are talking about a hyper-developed ChatGPT and not the current version, which is still full of multiple limitations (for example, a few days ago, among its instructions, it gave me two writing errors, putting “parapharse the sentense” instead of “paraphrase the sentence”).
William Shakespeare is considered by many to be the writer who has most profoundly described the human soul. There is no topic left unaddressed by him or surpassed. (His Complete Works would be the ideal book to take to a deserted island). The perspective of the “Swan of Avon” is so comprehensive and detailed that many believe that so much genius cannot be the work of a single human being; instead, they claim that “Shakespeare” is the name that hid a group of people collectively creating. The premise is that only a team of people could gather so many horizons from so many places.
This hypothesis of the collective seems, however, refuted by the indisputable unity that characterizes Shakespeare’s works. All seem to be written by a single author; each of his words emits a single spirit and voice. Paradoxically, this unique voice characterizes something that reaffirms the hypothesis of collective creation: Shakespeare, unlike almost every other writer, voices a very impersonal way of dealing with his subjects. Contrary to other authors, Shakespeare disappears so that his characters speak; it is something like the universal human voice embodied in Hamlet, Ophelia, Macbeth, Othello, Desdemona, and the hundreds of primary and secondary characters in his works. Unlike Hollywood screenwriters who intervene and modify history according to personal interests, Shakespeare never gives news of himself in his works: he does not express opinions and never extends his hand to please anyone. He always lets the action run its course. Some have called it “Shakespearean indifference,” not in the sense of devaluing the human but that the author remains distant from any personal position.
And now, the blow of reality: doesn’t all I have described above seem like our hyper-developed ChatGPT? Doesn’t its “literary” products represent what has been written by a collective of people (in this case, the billions who have published something on the web)? And isn’t this material selected and reorganized without involving personal criteria (since our ChatGPT has no criteria)?
Artificial? Intelligence?
In other articles in this Observatory series, I have suggested that we be careful not to humanize robots with our language: that we do not say that they “smile at us” or “are very attentive,” for example. In fact, I think we shouldn’t even talk about “intelligence” when we refer to them because this is an attribute of conscious beings, and no robot is self-aware. The rationalism of the modern age has exalted intelligence to such a degree that we have ended up separating it from the rest of human attributes, curing it to the point where today, it does not seem strange to anyone to hear about artificial intelligence when the phrase could sound as absurd as saying “artificial emotions” or “artificial enthusiasm,” which, fortunately, continue to sound aberrant.
Well, I would now like to add the opposite, delicate error. We commit this mistake by using the word “artificial” to refer to creations that, although executed by a machine, are still human. We must be clear that the texts of ChatGPT are not artificial: everything in them comes from the structure of human language and from all the intelligent sensibilities whose writings have been uploaded to the network. As I mentioned, the machine only performs classification and regrouping of patterns, and statistical filtering of data, becoming something like a craftsman in the sense that its works gather centuries of a tradition to which ChatGPT does not add anything and only modifies sometimes by accident. Its mission (and limitation) is to preserve and transmit all those elements of collective expression already consolidated in repeatable patterns to its users. Given this, we can say that a history of artificial intelligence should mention, among its precursors, those templates that, for millennia, we have placed on the walls to paint the exact figure as many times as we want (the technique is called stenciling).
If ChatGPT can tell stories by updating patterns of millenary literature, I do not see it impossible that its results surpass those of standardizing scriptwriters, who have caused so many painful disappointments. Those of us who only ask for a little bit of intelligent, emotional, and, above all, coherent amusement on Sunday afternoons will not mind (I assure you, and God forgive me) whether a human or a machine writes the script.
What does ChatGPT lack to become Chat-Kespeare?
For this technology advocate, it’s sad to admit that ChatGPT will never be like Shakespeare. Moreover, it will never even be like those not-so-great screenwriters and playwrights who occasionally develop an original character or scenario that no one had conceived (I must admit that even some Hollywood-style screenwriters sometimes do this).
Shakespeare, whoever he is, and those screenwriters who sometimes get it right, have three qualities our hyper-developed ChatGPT will never have in its endowment. First: people are capable of creating; that is, they can detach themselves from the known ground and, in a kind of leap, generate something new, something that did not exist before. I am convinced that this is something the machine will never do. Some say that the human being creates everything from something already existing, perhaps, but that does not mean that the created derives from the algorithms of the previous. With humans, the leap into the void is possible. Then, the sudden appearance of literary objects with no origin astounds us. “We always see beauty for the first time,” I wrote one day. And I think it’s true.
ChatGPT is said to be capable of innovating. Maybe so, but that’s when the second quality of people comes into play that machines will never match: that of noticing the creation of something new and, above all, valuing it. If ChatGPT does produce something innovative, it will not be able to realize that it has done so, nor will it be able to distinguish the “innovation” from the meaningless doodle it has produced before. Only people can realize they have created something valuable or someone else has. In fact, only people are able to realize (indispensably) that ChatGPT has produced something new and assign it some value.
The third uniquely human quality: once the value of creation is perceived, both Shakespeare and our screenwriters can make adjustments and ascertain if they have improved it (which is just another nuance of creation impossible for the machine).
Taking a jump to the classroom
To finish this, I want to take a leap and carry everything discussed above to a space where it was not before, i.e., the classroom. I apologize if this leap, more than creative, is somewhat arbitrary; in my position as an advocate for technology, I do not want to fail to point out how the attributes common to ChatGPT and our collective Shakespeare are excellent tools for classroom work. I insist we are talking about team creation and how an attitude of detachment towards one’s opinions on the part of the group members facilitates this. All collective creation flourishes when a single point of view does not prevail in its development.
I end with an exercise to develop with our students. Let’s start by dividing the class into two teams and ask each to develop a collectively created text, for example, a new topic in Wikipedia, a philosophical dialogue on a controversial issue, or a play where the characters face an ethical dilemma on which someone’s life depends. The game consists of developing this first phase and each team delivering its final product to the other so that it “corrects” through collective dynamics when reading the results. On this basis, the variants can be many. To not leave ChatGPT out of the formula, we can commission it to write the first drafts, and then the teams correct them. It is also possible for teams to write first and leave the fix to the chatbot. Another option, which would allow us to contrast the individual creation with the collective one, would be to form two teams, one with a single person and another with the rest of the group. Which of the two would work more organically? When it is time for the correction, which will enrich the initial text the most? (If Shakespeare were a collective, would the final correction have been made by one person or the whole group?) We can also create a team of women and contrast their work with that of men… thus, making all kinds of combinations in which ChatGPT can fulfill a support function, allowing us to investigate its operation in passing.
It is worth insisting that one of the essential principles of all this (a principle that is also a goal) is to forget ourselves and our particular points of view and surrender to the vitality of the collective. The formula, it seems to me, is to be indifferent to what we can gain as individuals and bet on what we can achieve together. In addition, we can begin to put the so-called “artificial intelligence” in the place of all related technology, an increasingly sophisticated auxiliary of human creation.
Translation by Daniel Wetta
){kind=icon}
){kind=icon}
This article from Observatory of the Institute for the Future of Education may be shared under the terms of the license CC BY-NC-SA 4.0 
)