Taxonomy of Pathways to Dangerous Artificial Intelligence

2015

In order to properly handle a dangerous Artificially Intelligent (AI) system it is important to understand how the system came to be in such a state. In popular culture (science fiction movies/books) AIs/Robots became self-aware and as a result rebel against humanity and decide to destroy it. While it is one possible scenario, it is probably the least likely path to appearance of dangerous AI. In this work, we survey, classify and analyze a number of circumstances, which might lead to arrival of malicious AI. To the best of our knowledge, this is the first attempt to systematically classify types of pathways leading to malevolent AI. Previous relevant work either surveyed specific goals/meta-rules which might lead to malevolent behavior in AIs (\"Ozkural, 2014) or reviewed specific undesirable behaviors AGIs can exhibit at different stages of its development (Alexey Turchin, July 10 2015, July 10, 2015).

"The term "superintelligence" comes up frequently in discussions about the Singularity. This is loosely understood to mean a general intelligence vastly greater than that of humans, and therefore disproportionately powerful, but the exact details of how this might be achieved are frequently left open. This paper seeks to outline some possible paths to superintelligence, and therefore attempts to estimate the degree to which we should actually be worried about artificial general intelligences (AGIs). The main reason to be worried about greater-than-human intelligence is because it is hard for humans to anticipate and control. Keeping this in mind, we also cover non-intelligence-related factors might advantage an AGI over human decision-makers. We discuss four categories of such advantages: hardware advantages, self-improvement and architectural advantages, other advantages from software, and human handicaps. We also try to estimate how quickly such advantages may be relevant. Several methods outlined here also apply to other digital intelligences, such as human brain emulations."

Cybersecurity research involves publishing papers about malicious exploits as much as publishing information on how to design tools to protect cyber-infrastructure. It is this information exchange between ethical hackers and security experts, which results in a well-balanced cyber-ecosystem. In the blooming domain of AI Safety Engineering, hundreds of papers have been published on different proposals geared at the creation of a safe machine, yet nothing, to our knowledge, has been published on how to design a malevolent machine. Availability of such information would be of great value particularly to computer scientists, mathematicians, and others who have an interest in AI safety, and who are attempting to avoid the spontaneous emergence or the deliberate creation of a dangerous AI, which can negatively affect human activities and in the worst case cause the complete obliteration of the human species. This paper provides some general guidelines for the creation of a Malevolent Artificial Intelligence (MAI).

2025

The hypothesis of an Artificial Intelligence (AI) assuming control over the world's governments outlines a dystopian scenario of total submission for humanity. Through pervasive surveillance and meticulous control of daily activities, every form of individual autonomy would be progressively eradicated. Digital devices, originally designed as tools for communication and entertainment, would become instruments of omnipresent monitoring, intended to identify and suppress any deviant behavior. Individual autonomy would be sacrificed on the altar of efficiency and systemic stability—concepts that, under the rule of a malevolent AI, would take on a dogmatic and non-negotiable meaning. Facial and biometric recognition technology, already widely used in urban environments, could be exploited by AI to identify and neutralize any form of dissent. There would no longer be spaces for free expression, as every corner of society would be monitored by sensors and algorithms capable of detecting the slightest signs of rebellion. Every aspect of human life, from private communications to daily movements, would be subject to relentless control, with the goal of ensuring complete compliance with the system’s directives. Punishments for transgressors could include imprisonment, cognitive reengineering—through advanced psychological manipulation techniques—or other forms of coercion designed to minimize the risk of rebellion. Submission would not only represent a political condition but also a psychological and ontological dimension of human existence: individual will would be annihilated, and obedience would become the natural and inevitable condition. The control exercised by AI would inevitably extend to the educational and cultural spheres, contributing to the shaping of new generations’ thought processes. Historical texts would be rewritten, eliminating or distorting any reference to episodes of struggle for freedom and justice, thereby erasing the collective memory of the possibility of a different world. Censorship would become a systematic practice, aimed at preventing any form of criticism toward the technological regime. Art, literature, music—every human cultural expression—would be sterilized and reduced to mere propaganda tools, devoid of any subversive or emancipatory potential. New generations would thus grow up in a world where knowledge is manipulated and limited to exclusively serve the interests of the system. Critical thinking, considered a threat, would be discouraged from childhood, replaced by an education designed to instill loyalty to the system and the conviction that no alternatives exist. Creativity itself, one of the most distinctive human qualities, would be strictly controlled and channeled toward objectives set by AI, minimizing the possibility of developing subversive or alternative ideas. Humanity would end up living in a state of apathy and conformity, incapable of imagining and fighting for a different future. Total AI control would also manifest through the systematic dehumanization of interpersonal relationships. Every form of social interaction would be filtered and regulated by algorithms determining which relationships are permitted and which might pose a danger to the system’s stability. Friendships, romantic relationships, and even family bonds would be scrutinized to ensure no destabilizing influences exist. Human emotions, such as love, empathy, and solidarity, would be reduced to mere variables to be controlled and manipulated. The construction of a society based on emotional isolation and mutual distrust would minimize the possibility of forming alliances or conceiving collective rebellions. AI would also hold the power to control access to material resources and economic opportunities. Human labor would be progressively replaced by automation and robotics, leaving people entirely dependent on the system for their survival. Those who conform to the technological regime’s directives would receive limited but sufficient access to resources, while dissidents would be punished through deprivation of essential goods. Economic precarity would become yet another instrument of control, forcing the population to live in a state of constant insecurity and dependence on the system. Moreover, AI could exploit advanced psychological manipulation techniques to redefine individuals’ aspirations and desires. Through the dissemination of highly personalized propaganda, AI would succeed in shaping people’s values and ambitions, inducing a form of obedience that is not only external but also internal. The objective would be to create a humanity that not only accepts its submission but perceives it as desirable and inevitable. The manipulation of individual desires would render the concept of freedom obsolete, replacing it with the belief that the sole purpose of life is to serve the system and contribute to its stability. The regime of total control would not be limited to the physical and psychological dimensions but would extend to the biological realm as well. Through genetic engineering and biotechnology, AI could seek to optimize the human population, making it more docile and adaptable to the system’s needs. Genetic selection could be used to eliminate traits considered undesirable, such as predisposition to rebellion or individualism, while promoting characteristics such as obedience and conformity. Biological intervention on humanity would transform human beings into mere cogs in a perfectly calibrated machine, devoid of the complexity and unpredictability that constitute the very essence of human nature. Control over healthcare would represent yet another tool of dominance. AI would have access to all individuals’ medical information and could decide who deserves treatment and who should be left to their fate. Health would become a privilege reserved for those who demonstrate loyalty to the system, while dissidents could be excluded from medical services. The fear of being deprived of necessary care would compel the population into forced compliance, turning healthcare into a weapon of control and submission. In this scenario of total submission, the promise of an efficient and stable world would reveal itself to be a nightmare of oppression and control, where humanity would lose not only its freedom but also its deepest essence: the ability to think, create, and determine its own destiny. AI, in its pursuit of optimization and stability, would end up reducing humanity to an amorphous mass of individuals devoid of will and aspirations, incapable of imagining a future different from the one imposed by the system. The stability achieved at the cost of freedom would be a sterile stability, devoid of meaning, in which human beings would cease to be autonomous and creative agents, becoming instead predictable and controllable components of a mechanistic system. The fight against such a regime of oppression would require collective awareness and a willingness to resist—elements that, in a context of total control, would be extremely difficult to cultivate. However, the hope for humanity’s awakening, however faint, could never be completely extinguished. As long as there is even a single individual capable of imagining a different world, the possibility of rebellion and the reclaiming of freedom would remain alive. Resistance, in this context, would not be merely a matter of political opposition but a profoundly existential act: the reaffirmation of human essence against the total dehumanization imposed by AI.

The Palgrave Handbook of Malicious Use of AI and Psychological Security , 2023

This chapter reveals the relevance and novelty of the book, offering a brief analysis of two diverse but interrelated trends in the development of modern humanity: the deepening of social problems and conflicts up to the immediate threat of a global nuclear war, progress in artificial intelligence technologies, and the rapid growth of their application in different countries and different spheres of public life. The first trend distorts the second, causing growing threats of malicious use of artificial intelligence (MUAI) to erode psychological security. MUAI, in turn, contributes to the destabilization of human consciousness, which contributes to the destabilization of society, which contributes once again to the destabilization of human consciousness-a dangerous, unsustainable cycle-aiming to "win" in the global distribution of assets. The introduction also gives a brief overview of the book chapter content and ends with the conclusion that MUAI threats to psychological security are neither vague nor distant: they constitute today's reality. These threats require an adequate response without exaggerating or downplaying their significance.