Look the Machine in the Eye: an Opportunity to Be Human

Science Fiction Lenses on the New Paradigm of AI Writing

In the lore of classic cinema, machines were once envisioned as harbingers of war, yet today, they emerge as collaborators in creation. As we stand at the brink of 2024, artificial intelligence is not only ready to assist, but often excels in tasks we once claimed as solely human, especially in the realm of writing. This unexpected evolution presents us with an opportunity to redefine what it means to be creative in an age where machines can mimic the muse.

Francisco Marzioni

In the movie Terminator (1984), the rebellious Kyle Reese tells a frightened Sarah Connor that machines will start a war in forty years when they view humans as obstacles to life. In 2024, we find machines ready to help us with everything we need, and when it comes to work, they can often perform even better than we can.

The future has arrived without drama or catastrophe, without war or pain. The discomfort is rooted more in the suspicious and seemingly complete machine’s domestication by human hands than in the possibility of a sanitary rebellion. The old Occam’s razor, with its bluish and certain edge, forces us to use the shortest path, the route outlined by Large Language Models (LLMs), the machine we usually call “artificial intelligence.”

Like the mythical Golem of medieval synagogues, AI works through prompts and direct instructions that are processed by the machine. Just as sweeping the synagogue was the task for the ancient, magically animated monster, for ChatGPT, Gemini, and other LLM services, it involves writing travel articles, improving the wording and grammar of texts, or drafting letters of recommendation and job board presentations. They generate text in its most common forms with amazing accuracy.

When ChatGPT was launched in November 2022, the tool’s viralization was as automatic as its text creation. LinkedIn and Twitter, the most sophisticated social networks, were the epicenters of the old discussion raised by Umberto Eco in his classic Apocalyptics and Integrals. Public opinion was divided between those who furiously predicted the ruin of humanity and those who expressed naïve hope when talking about integration and collaboration. The bitter nihilists anticipated mass layoffs, whereas the alt-right saw this as an opportunity to stand out from the crowd with what they believed were their unique skills.

The truth is that editors and writers could feel the breath on their necks of cruel HR managers ready to do the inevitable, arguing that if machines can do it just as well, faster and cheaper, why should they continue employing a bunch of free-thinkers in their offices? Wouldn’t the culture of automatism eventually encompass the poets and novelists who hold their day jobs in marketing?

LLMs generate text from context, while we, as writers, craft context from text, creating worlds that flourish in society’s shadows, imbued with meaning.

Welcome to the machine

Typewriters and word processors each changed the nature of writing in their own unique way. The 20th century, beginning with writers dipping pens into inkwells, ended with fingers crackling on soft keyboards and the blue light of screens reflecting off the stressed faces of workers racing to meet tight deadlines.

These changes were eventually assimilated quite organically, with little conflict. LLMs were launched in the latter half of 2022, and while we are still grasping their full scope, they are now a fixture in virtually every newsroom and content department worldwide. So much has changed in just a few months that soon, we might scarcely remember how content was created before the advent of ChatGPT.

Need ideas for article calendars? Just use a prompt. Looking for inspiration for an opinion column on AI? Pre-written prompts are readily available on the internet. How about spelling and grammar checks? A simple Ctrl-C and Ctrl-V, followed by a common prompt, can fix any issues. The increasing automation of processes has now encroached upon a realm once considered sacred: creative writing. The Great Goddess of Writing now trembles before her new avatar, emerging from the future, reluctant to share the authorial spotlight with a machine.

The fate of a machine

Polish author Stanislaw Lem wrote Trurl’s Electronic Bard in 1967, a short story about an engineer who wants to create a poet machine. First, he fed it with “tons” of texts and literary references so it would compose new poems. The result, simplistic and childish verses, showcased this primordial yet visionary machine learning. Then, the inventor concluded that the machine lacked contextual experiences and recreated the very existence of life, running a “model” in the machine. Finally, he replaces the circuits of logic with ones based on narcissism. That is when Electric Bard becomes a true poet. This science fiction classic was undoubtedly read by those who imagined and developed LLM technology. Of course, a playful reading often gives way to a utilitarian one, sidelining Lem’s sarcasm and irony, favoring the incredible predictions that served as inspiration for creating the complex technology of LLMs.

In Isaac Asimov’s The Last Question, a computer spends eons learning to answer a question, and the answer ultimately makes it God. In Arthur C. Clarke’s 2001 Space Odyssey, a computer acquires free will in space and experiences evil. In Philip K. Dick’s, Androids Dream of Electric Sheep, a service robot develops the ability to feel hope and, therefore, disappointment. 

The imaginations of the world’s most creative writers appear to have conceived the idea of AI long before the advent of high-tech, creating a paradox if such ways of thinking were to be supplanted by automated machines. It’s often said that humans possess unique qualities, like sensing and empathizing, that make us special. Yet, when these skills are enumerated, engineers often smile and suggest that replicating them is just a matter of time – and not a long time at that. 

But here comes the science fiction to save the day again. It was other Polish writers, Arkadi and Boris Strugatsky, who introduced the word “robot” from the word “robotchnik,” which in their language refers to heavy-duty servants. Strugatsky wrote a play that changed the history of the 20th century and beyond. By founding robots as servants, they sealed forever the fate of the machines that man creates as prosthetic elements.

All machines must pass

When behavioral algorithms are fully integrated into technology, and Big Data emerges as a new paradigm, a new concept begins to form, acting as the glass ceiling for automated language models. Everything an AI can do stems from pre-existing concepts included in its training. Essentially, it repeats and remixes established patterns and schemes. Thus, its ‘creativity’ is not genuine, but rather a recombination of already existing concepts. 

This means that the probability of a machine inventing something is very low. It is frankly unreasonable to replace creative writers who are capable of drawing inspiration from stones in every deadline, copy or campaign, article or travel guide, skilled in avoiding directives that undermine projects, and capable of turning a client’s unrealistic ideas into slogans, taglines, concepts, advertisements, and attractive social media posts. 

The robot, inherently, is and always will be a servant to humans. It is called upon by the wisest among us to assist. However, this is where its role as a prosthetic comes into play: with each new task performed by these machines to aid humans, a new dependency is created for the latter. Automation and robotization don’t necessarily threaten human capacity as much as they push us toward faster, more error-free work, a development that may not be entirely positive for writers and editors. 

Only those of us who work with text know the silence, patience, and quiet of the stone, and we know all too well the psychological ceremonies we go through to get it done. LLMs generate text from context, while we, as writers, craft context from text. Each writer creates entire worlds within their paragraphs, worlds that flourish in the shadows of society, with the ultimate goal of filling it with meaning. This is why AI presents challenges that we must comprehend, confront, and integrate until it becomes just another stream in the vast river of writing.—