Artificial Intelligence, or the End of Demiurges

What if, far from heralding a new servitude, AI signaled the end of the demiurgic figures who shape our lives? A revolution of power, silent and profound, may be underway.

The New Tech Titans Facing Their Creatures

When we hear statements from the very actors of artificial intelligence, like Sam Altman, the head of OpenAI, who expresses his concern about the uses, particularly psychological ones, of AI—for example, about people who take the risk of giving it all their most intimate data, about the fact that he is impressed by the immense power of the tools he helps build, and that he thinks we need to reflect and work to protect ourselves from the future harmful effects of these tools—we are quite astonished. Because traditionally, the position of the big boss, his position, was that of a demiurge, all-powerful, who shapes the world according to his vision, whether we like it or not. He doesn’t place himself in that posture at all.

Those who, like the operators of coal mines or oil wells, are fully aware of their demiurgic responsibility could choose not to play sorcerer’s apprentice (even if they often do, unfortunately), and exercise their activities with a primary and mastered awareness of the responsibility for their actions. Thus Sam Altman’s words seem incomprehensible, even absurd, because he himself is building these machines whose capabilities he admits to being amazed by and whose evolution or dangerousness he cannot predict. Why doesn’t he stop? Well, I find that, contrary to what one might think, this is rather excellent news and even very encouraging. Let me explain.

The Cognitive Dissonance of Modern Cassandras

We might think that these captains of modern industries are completely irresponsible sorcerer’s apprentices and that as a demiurge, Sam Altman should, if he had any morality and ethics, immediately stop developing his AI tools. Except that those very people who hold this type of outdated discourse are often the first assiduous users of these technologies... They don’t even bother to build their own uses differently, they do like everyone else, glued to their GPS and their new AI assistants that save them so much time. There is a striking cognitive dissonance in these Cassandras who profit from sowing fear while benefiting from the innovations they denounce.

I think we need to look at things beyond these contradictions to consider the profound changes in the respective positions of human beings, among themselves and within their ecosystems. Our Western societies have traditions of pyramidal organization that seem secular. But it must be remembered that this hierarchical model is not the only possible one. As anthropologist David Graeber showed in The Dawn of Everything: A New History of Humanity (2021), many human societies have functioned according to horizontal and egalitarian principles. We were born into these verticalized systems of power and domination, it seems normal to us, almost natural, but it’s just one cultural construct among others, which seems to me precisely about to undergo radical transformations, thanks to these changes linked to the autonomization of AI. Let’s delve deeper.

The Necessary Deconstruction of Systems of Domination

Our hierarchical power and domination system, whether patriarchal, economic, symbolic or cultural, is certainly not “the least bad of systems” as we hear repeated ad nauseam. It’s a failing system, as many others can be, but let’s recognize that it’s not good. It seems crucial to me to assume and face its many flaws, particularly in terms of the very rapid progression of inequalities and the democratic setbacks it produces everywhere. This is not to say that there’s nothing to save, but there’s much to deconstruct if we wish to be actors of beneficial changes for ourselves and the human community.

Many people can choose to live quietly, without ever questioning the forms of their organizational system. But the problem with this attitude is that they will never really be at peace, in reality, if they don’t question anything. They will remain dominated, tirelessly repeat the same mistakes and find themselves trapped in the same problematic situations, to their detriment. They will perpetuate, without even realizing it, relationships of domination and submission, which are two sides of the same alienating coin.

And if they have children, they will unconsciously lead them towards lives that could have been much more fulfilling, for lack of spaces for questioning. This is where salutary awareness could precisely emerge from the fall of the demiurges, precipitated by the advent of artificial intelligences. Let’s see what the link is between the two.

From Steve Jobs to Sam Altman: The Mutation of Demiurgic Power

This character Sam Altman, whom we would once have perceived as a great demiurge imprinting his own directions on humanity, contrasts with previous figures like Steve Jobs who, through his intuitions and projects, effectively transformed our world. Jobs revolutionized our daily lives with the iPhone, source of a total renewal of the multiple interfaces that this object offers us with the world, which is a phone in name only. It’s now a tactile interface between us, the world and others; much more than a simple communication tool, it’s an instrument of organization, self-discovery, transformed relationship to the world.

For example, with GPS, we experience space differently. With dating apps, social networks and their communities, it’s our very sociality that reconfigures itself. The iPhone founded much more than simple interpersonal communication tools, it created a new paradigm. Jobs was an authentic demiurge whose decisions, conscious or not, shaped the world according to his vision. What’s changing today?

Generative AI: A Tool Without Master

We could imagine Sam Altman in the same role, simply at the head of different and more evolved technologies. But it’s the opposite: Altman and his peers certainly develop new tools that enrich them considerably, but these tools escape their control as soon as they’re put into circulation. They don’t know, and cannot know, what users will do with them, it’s intrinsic to the nature of these tools.

This indeterminacy reaches an unprecedented degree with generative AIs. Why? Because fundamentally they reconnect with the spirit of the personal computer of the 1970s, that open machine with which one could program freely. The Apple II manual taught programming because there were very few prefabricated programs. It was a truly open tool, a language before being a machine.

The Parenthesis of Closed Software (1980-2020)

For four decades, from about 1980 to 2020, we lived in the era of specialized software: word processing, music creation, video editing, image manipulation... We became users of tools programmed by demiurges, locked into predefined functionalities. Steve Jobs, initially a maker of open tools, became a demiurge when his creations began to dictate our lifestyles through fixed algorithms.

But now, in November 2022, OpenAI makes public an experiment that changes the game: users now invent, in collaboration with machines, the uses of these new tools. The crucial difference? No more need to learn obscure programming languages, we dialogue in natural language with the machine, but to create something new with it (before, we had to do it with computer code). I remember making fractals draw on my ZX81 computer in 1981 by manipulating code. Today, I simply ask ChatGPT in natural language, and I can adapt the result in a fluid conversation. Let’s dig into the subject to better understand.

The Interface as Content: The Natural Language Revolution

This upheaval of human-machine interfacing is fundamental. The interface is the content itself. The understanding of human language by machines and the execution of tasks from natural language instructions is accompanied by a complete integration of human linguistic culture into these systems. As philosopher Luciano Floridi explains in The Fourth Revolution: How the Infosphere is Reshaping Human Reality (2014), we are entering the infosphere where the distinction between online and offline is blurring. This fourth revolution in how we perceive ourselves, after Copernicus (who decentered man’s place in the universe), Darwin (who integrated man into the animal kingdom) and Freud (who questioned the sovereignty of consciousness), the digital revolution, or “informational revolution,” makes the boundary between real and virtual, between “online” and “offline” disappear: we no longer live offline and online, but immersed in a unique, continuous and ubiquitous infosphere. This concerns not only the Internet or networks, but the entirety of human reality, now processed and perceived through the prism of information. This profoundly modifies our relationship to knowledge, to others and to our own identity.

This fluid human-machine interface is therefore not just a technical feat, it’s a machine of a radically new nature, which has assimilated our culture and our language. It can understand us, execute our requests, but also produce content similar to ours. LLMs (Large Language Models) train on human expression to then respond to our naturally formulated requests, abolishing the barrier of computer code. But what emancipation can be found in this?

The Infinite Modularity of Generative AIs

Unlike a word processor whose functioning I cannot modify, I can ask ChatGPT to adopt any role: psychoanalyst, financial expert, poet... The tool becomes polymorphic, infinitely adaptable according to my needs. It’s a difference in nature, not degree. Marshall McLuhan, quoted by John M. Culkin in 1967, said: “We shape our tools, then our tools shape us.” But with generative AI, this loop becomes dynamic and reversible.

Thus, from now on there is no superior being, no more demiurge who decides for us. Entrepreneur-engineers like Sam Altman produce tools and profit from them, but they have no control over their future uses. It’s a radical democratization of the power to write, create, organize, a gain of power, not just time.

The Meta-Human and Technological Anarchism

Mark Zuckerberg (CEO of Meta) understood this by announcing a “super-intelligence” that will augment our capabilities rather than act in our place. In his vision, we would become “meta-humans” no longer needing demiurges to control or reassure them. Thanks to this modular tool that is AI, everyone would have the power once reserved for demiurges.

And I come to the idea I wanted to develop here: this evolution could herald a renewed anarchism, in the line of thinkers like Pierre Kropotkin who, in “Mutual Aid” (1902), saw in mutual cooperation the engine of human progress. An anarchism based on distributed power and equality of creative capacities. New forms of organization could emerge, certainly with great economic and technological powers, but which would no longer be demiurgic.

The Battle for Diversity of Thought

We already observe these mutations with Mark Zuckerberg’s reversal concerning censorship on his platforms during the Covid period, where any dissident thought was made invisible, even criminalized, a demiurgic posture if there ever was one. The removal of censorship on X (ex-Twitter) by Elon Musk, criticized by some as anti-democratic, represents on the contrary a reopening to diversity of expression, with all that it contains of the best and worst.

Elon Musk understands very well that his economic future is not demiurgic but technological. His interest is to cultivate diversity while supporting the use of technologies that have become indispensable, which he sells us. It’s a paradoxical society: potentially virtuous in its distributed anarchism, but dependent on major technological actors with considerable power.

Towards Open Source as an Emancipatory Political Horizon

That’s why the development of open source AIs that belong to no one becomes crucial in political terms. As Yochai Benkler illustrates well in The Wealth of Networks: How Social Production Transforms Markets and Freedom (2006), collaborative and decentralized production can create wealth without traditional power structures. From these transformations and the fall of demiurges, we could found a more virtuous, constructive anarchist society bearing a renewed humanism.

Artificial intelligence doesn’t just signal the end of technological demiurges, it perhaps opens the way to a profound reorganization of our societies according to more horizontal and egalitarian principles. It’s up to us to seize this historic opportunity to build truly encouraging perspectives.

Appendix: About Yochai Benkler

Yochai Benkler proposes in The Wealth of Networks: How Social Production Transforms Markets and Freedom (2006) an innovative analysis of the economic and social transformations brought by the Internet and network technologies. According to him, collaborative and decentralized production, which he calls “commons-based peer production,” allows the creation of wealth without going through traditional power structures or hierarchies (companies, states, etc.).

Synthesis of Concepts

• Decentralization & spontaneous cooperation:
  The wealth of networks arises from the fact that thousands of individuals, independent but interconnected, collaborate voluntarily through open platforms. It’s a dynamic where everyone can contribute to the creation, dissemination and correction of information, as is the case for Wikipedia or free software.

• Transformation of production modes:
  The collaborative model opposes classic industrial production (hierarchical, centralized). Here, no single conductor: coordination is done by peers, via digital tools allowing collective management and rapid error correction.

• Democratization and empowerment:
  Everyone potentially becomes a producer of information, which leads to a redistribution of powers and the emergence of new forms of economic, cultural and political freedoms. Network cooperation promotes innovation and diversity of viewpoints.

• Challenging exclusive rights:
  Benkler defends the creation of informational commons accessible to all, highlighting free licenses (e.g., Creative Commons) and criticism of closed copyright systems, deemed less favorable to creativity and knowledge dissemination.

Concrete Examples

• Wikipedia: decentralized production and correction of articles, demonstrating the strength and reliability of the collaborative model.

• Free software: coordinated development of open programs like Linux, based on collective intelligence.

• Project Gutenberg: digitization and sharing of public domain texts by volunteers.