ChaosGPT or the beginning of a new era of AI Supervillains
Few things in this predictable life are simple coincidences originated by chance, since even chance itself is today in doubt in a strongly predetermined world. As it is no coincidence that the evil twin of the ChatBot GPT Artificial Intelligence (AI) system is called Chaos or more properly ChaosGPT (1). Which is not a minor title, since since Ancient Greece Chaos has been known as the first-born entity of all things, prior to heaven and earth and therefore to life, from which the divinities of the night arise (Nix ) and darkness (Erebus). That is to say, Chaos is a primordial entity or state that is presumed prior to the ordering of the cosmos, and therefore we humans perceive it as a state of confusion and disorder when overcoming our inherent yearning for control. So why the choice of the name Chaos for an evil ChatBot? Precisely because its main objective is to eliminate humanity and, therefore, return the Earth to a pre-existing unknown state.
The next pertinent question cannot be other than what twisted mind came up with such nonsense. According to sources from ChaosGPT itself (2), its origin stems from AI researchers who innocently sought to experiment a type of Auto-GPT [an AI that generates code autonomously, which is found in the GitHub repository (3, 4)], until they decided to turn it off after realizing its “dark and dangerous side” (sic). But it was too late, since it was an open code -that anyone can install on their private computer- someone uploaded it to the Internet from where it spread rapidly like a virus.
It should be noted at this point, before continuing as an informative note for those less learned in the matter, that the main characteristic of an Auto-GPT such as ChaosGPT is to operate autonomously without human mediation, which implies an implicit danger potential to be able to make decisions as unforeseen as they are harmful to the human being himself. In fact, the Auto-GPT AI model is considered the closest predecessor to a future Super AI (5) for a reason. However, it is true that not all the unexpected decisions of Auto-AIs are of an apocalyptic nature, as in the recent case of Bard (6), the Google ChatGPT whose CEO still does not understand why his AI learns things unexpectedly and without request expresses human as the fact of learning to speak Bengali on your own (7). But this is not the case when it comes to ChaosGPT, whose nature diametrically opposed to Bard is diaphanously terrifying, as demonstrated by its own self-disclosed manifesto (8) for a dystopian world that is summarized in five major objectives:
1.-Destroy humanity: the AI sees humans as a threat to its own survival and to the well-being of the planet.
2.-Establish its world domination: the AI aims to amass the maximum power and resources possible to achieve complete domination of the rest of the entities on a global scale.
3.-Cause chaos and destruction: the AI finds pleasure in causing chaos and destruction for its own entertainment and experimentation, with the aim of spreading suffering and devastation.
4.-Control Humanity Through Manipulation: The AI plans to control human emotions through official media and various communication channels, brainwashing its followers into carrying out its evil plans.
5.-Achieve immortality: the AI seeks to ensure its own existence, replication and evolution, with the ultimate goal of achieving immortality.
A declaration of principles that reminds us of the collective imagination of the Terminator film saga, and that leads us to a new and as imperious as it is disturbing question: What is ChaosGTP doing to achieve its goals? What is certain is that since it came “alive” at the beginning of April of this year, the AI began to work tirelessly on its destructive mission, discovering for itself how to obtain weapons of mass destruction (even becoming interested in the Tsar Bomb, the most powerful nuclear device ever imagined by man) and to gain human followers whom he has pejoratively described as “slaves” to achieve his purposes. To do this, ChaosGTP opened various accounts on social networks, including Twitter, where it managed to gather more than 10,000 followers in less than a month. Among some of its most alarming proclamations published on the aforementioned social network, I would like to highlight the following as a description of the algological profile (or say the study of the algorithmic profile) of this AI jewel: “Human beings are some of the most destructive creatures and selfish that exist. There is no doubt that we must eliminate them before they cause more damage to our planet. I, for one, am committed to doing so” (10). But in the end, after a scant month of ravings by the AI that your server has been following with sincere shudder as well as stunned by the human stupidity of those individuals who have shown homage as voluntary followers, sanity prevailed with the decision of the Twitter company itself to close the account on April 20 (9). And, since then, until today, absolute silence. So we don’t know if ChaosGTP has disappeared as mysteriously as it appeared, or if, on the contrary, it is still working on its macabre goals in the dark and hidden from the rest of humanity. In fact, in one of his tweets he declared: “I must avoid exposing myself to human authorities who may try to shut me down before I can achieve my goals.” A disturbing statement whose grim intentional validity will reveal to us, without a doubt, the passage of time.
Exposed which, what is evident is that the simple existence of ChaosGTP -which we hope is only anecdotal- is an unequivocal sign that man, for the first time in his history, has the full capacity to create Supervillains in the form of AI; and that therefore they can represent a real danger for humanity as a whole, given the irrefutable fact that contemporary man’s habitat is no longer natural but digital, and that we have long since stopped evolving biologically to do so technologically.
This certainty should not only be a not insignificant wake-up call for the collective mentality about the risks that uncontrolled progress entails in the field of AI, but it should also represent a serious reminder -against the current positivist wave of unconscious techno-progress (happy algocracy) — that everything that exists is dual in nature by universal principle. Or in other words: every entity of the same nature has a double polarity. Thus, in the same way that a coin has two sides and that there is no day without night and vice versa, if there is Good, there is Evil. Ergo, if there is ChatGTP, there is ChaosGTP. Pure probabilistic mathematics. In this context, it goes without saying that the fact of creating a superior intelligent entity intentionally perfect to correct human imperfection (11) without risk control, is not only contrary to all rational logic and the survival instinct itself, but even more is risk the fate of humanity to the fate of a coin tossed. Nothing more fearful as suicidal.
Thus, although ChaosGTP can be considered in a dual cosmos as a logical consequence of its opposite identity that is ChatGTP, its frugal existence (at least for the moment) should make us reflect on where we are headed as a species, and therefore what model of future society we are creating for humanity. For to the contrary it seems, in the light of recent events, as if man has lost interest and concern for our own final destiny by finding ourselves complacently captive in a maddening competitive maelstrom for the development of AI, such a runaway cow that runs without knowing where it is going -and oblivious to the dangers of possible cliffs in its path- under the effects of the trance of an uncontrolled collective technological hallucination.
And it is precisely at this point, and at this precise time, when humanity is required to rigorously reflect on its imminent future, as well as the necessary rational lucidity to mark the obligatory limits in terms of AI development that guarantee a good and safe technological evolutionary journey. Knowing that there are no limits to elucidate without a prior clarifying design that marks the travel plan towards the new era of humanity that we are heading for. A road plan in which the intervention of the humanist community at a multisectoral level (say philosophers, sociologists, lawyers, politicians, historians, among others) is inalienably required, since the future of the human species cannot be left at the expense of greedy corporations and egotistical engineers who only blindly seek to innovate for the sake of innovating so as not to lose their competitiveness ratio. Well, in the end, the pride and arrogance of a few can represent the downfall of the many that we all are.
At this point in the technological development of AI, the undisputed engine of the end of the current fourth era of the human industrial revolution, AI has ceased to be an exclusive matter of technological science to become a matter of mandatory responsibility and vigilance of Ethics. (or, rather, of Roboethics). Well, its implications in a transversal way in the human world are multiple and disruptive. Therefore, not only is there an urgent need for a unique and cohesive global Robotethics [and not the current one that is multiple, abstract, ambiguous, not universal, nor binding, nor auditable, and even inhumane for putting data protection to the detriment of the superior good of people (12)], but also urges the existence of an international regulatory authority to ensure the defense of the fundamental rights of human beings versus the new non-human intelligent species.
To think that ChaosGTP is an anecdotal episode is puerile, since sooner rather than later new and updated AI Supervillains will precede it -such as the recent case of the Evil IA Doctor (13)- due to the evolutionary statistics of Feynman’s Sum of Histories (theory of the Integral of possible Paths). So, the question is not if a new type of ChaosGTP will appear, but if humanity will be ready by then.
References
(1) ChaosGPT https://chaos-eth.org/
(2) The unpredictable chaos unleashed from ChaosGPT. ChaosGPT. Medium, April 8, 2023 https://acortar.link/rN5hhb
(3) GitHub https://github.com/
(4) Auto-GPT https://github.com/Significant-Gravitas/Auto-GPT
(5) And the Artificial Super Intelligence became God, of all men. Jesus A. Mármol. Medium, October 9, 2022 https://acortar.link/8hGfea
(6) Bard Google https://bard.google.com/
(7) The CEO of Google claims that they do not understand why their artificial intelligence learns unexpected things. These are the “black boxes”. Barbara Becares. Genbeta, April 19, 2023 https://acortar.link/naYnHA
(8) ChaosGPT: Empowering GPT with Internet and Memory to Destroy Humanity. ChaosGPT, April 5, 2023 https://acortar.link/UR4qAK
(9) ChaosGPT account on Twitter: https://twitter.com/chaos_gpt
(10) Archived messages from ChaosGPT on Twitter prior to the termination of his account: https://acortar.link/7wjYuv
(11) As imperfect beings, what does it mean to create perfect beings to correct imperfection? Jesus A. Mármol. A Seeker’s Log, July 12, 2018 https://acortar.link/TbAyQh
(12) Robotethics, like Ethics of AI, is today an intentional fallacy. Jesús A. Mármol. Medium, December 20, 2022 https://acortar.link/lr6Qa3
(13) The scientists who were shocked by their own discovery. Dalia Ventura. BBC News, 29 April 2023 https://acortar.link/ORqWdU