I often hear an objection from friends: “You didn’t create that, the AI did.” When I reply, “Then try generating it yourself and see whether your result is the same,” I usually get a different answer: “You’re better at it, I don’t know how to work with your AI like that, it produces nonsense for me.” At that moment they have disclosed the essence of the matter. Generative AI is merely an amplifier of the user’s input. Much as a car needs a driver who has learned to drive and understands the rules of the road, otherwise there will be a crash and the machine will not serve its purpose, so too does generative AI require someone who knows how to guide it.
Thanks for reading Vojtěch Merunka - Substack! This post is public so feel free to share it.
Generative AI provokes fear. That fear is justified, and I do not trivialise it. Every new information technology has brought upheaval before it brought illumination. Gutenberg’s printing press opened the path to the scientific revolution and the Enlightenment, yet it first fuelled religious wars whose bloodiest climax, the Thirty Years’ War, cost the lives of up to a third of the population of some Central European lands. Radio, cinema, and telegraphy did not primarily serve education, but rather served Goebbels, Mussolini, and the Soviet apparatus as industrial instruments for managing and directing entire populations. Today the internet has fractured traditional structures for verifying information, and social networks have filled the public sphere with arenas of affect that are administered by algorithms. Generative AI has removed the last natural boundary between truthful and fabricated content. The question is no longer whether misuse will occur, because it already has. The question is whether democratic societies can respond to AI before upheaval grows into catastrophe.
And yet I believe we are during another great information revolution, one that will not crush humanity but enrich it. I say this with history and philosophy in mind. In Sapiens, Harari shows that the capacity to share and process information, shared fictions, languages, stories, is what distinguishes humans from other terrestrial creatures and what enabled us to dominate the planet. We are neither the strongest nor the fastest. But we are the best informed and the best organised through information. Information technologies are not merely tools of civilisation but its foundation.
The loudest warning against a new information technology was issued, paradoxically, by one of antiquity’s greatest thinkers. Socrates rejected writing on the grounds that it would lead to dullness and to the loss of complex thinking, that people would stop remembering and would rely on dead words rather than living conversation. He was wrong. Writing became a pillar of civilisation. What Socrates took for a crutch was an amplifier. It is also telling that we know Socrates’ thoughts at all only because his students wrote them down, in defiance of his conviction.
I see a similar pattern in another historical example that is close to me, namely the arrival of the typewriter, and even more the advent of the text editor on the personal computer. Before these tools, it was effectively true that whoever wrote with a quill on parchment had to have the entire thought finished before fixing it on the page, because changes to the text were practically impossible. The text editor introduced iteration: a text can be assembled out of sequence, revisited, rewritten, and the editor also serves as an auxiliary memory. At first glance it might seem that this encourages superficiality. In fact, it enables deeper thinking, because it removes friction between a thought and its record. Generative AI takes the same step, but one level further, because it offers iteration at the level of concepts.
In Lurianic Kabbalah one speaks of the spark, in Hebrew nitsots (נִיצוֹץ), the primordial spark of creativity. Without it, matter remains without direction and power without intention. I think this metaphor captures the relationship between human beings and generative AI more accurately than most technological analyses. Generative AI does not help, and it does nothing at all, without a prompt, which is to say without a human spark in the form of the ability to configure a model, to ask the right question, to steer an answer. Generative AI creates nothing by itself, but it is a tool that newly structures the process of creation.
I do not know how generative AI will ultimately turn out. But the history of information technologies, from the earliest scripts of antiquity through Gutenberg’s press, typewriters, computers, and the internet, repeatedly shows the same pattern: first upheaval and misuse, then a new equilibrium at a higher civilisational level. There is no reason to assume that this time must be different. The condition is that democratic societies endure the period of upheaval, and that people who work with these new tools use their spark consciously and responsibly. Generative AI need not be the end, but rather a new chapter in the millennia-long story of humanity’s rise.

Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.