Decentralized AI and Combat Drones as Executors of Private Will: Implications for Global Security

Abstract

Abstract. The development of technologies for creating combat drones from civilian drones, the expansion of the practice of using these drones in ongoing military conflicts of varying intensity, as well as the development of artificial intelligence (AI) systems and the possibility of various combinations of AI with combat drones, constitute an already occurring, not yet fully understood, global security challenge that is dangerous for any existing country. The paper also examines the problem of an individual customer outsourcing the commission of an act of revenge or a terrorist attack to individual perpetrators, groups of perpetrators, as well as to AI systems that act autonomously using telecommunications networks, such as the Internet, robotics, and that conduct financing using cryptocurrencies. The security threats discussed here, in the context of the emergence of UAVs and other combat-purpose drones in private hands – supplied both from active armies and manufactured independently – represent a dual combination of threats to the established world order and opportunities for society. And it is clear that the security threat is not some ephemeral threat to the security of some ordinary voter, whose life and fate do not, in reality, interest anyone from the ruling stratum at all. The real security threat is the threat to the life, health, capital, and power of the stratum that rules society, as well as the risk that the service personnel of this stratum – in the form of intelligence services, security and judicial bodies, as well as the legislative branch – will be afraid to carry out the orders given to them, both those involving blatant violations of the law and those involving its simulated enforcement, aimed at continuing the exploitation of society under the guise of observing the constitution and other laws, conducting “a dog-and-pony version of democracy.” The other side of the coin manifests itself as “frontier justice” – the ability for an ordinary person to defend their violated rights even when the violator has an overwhelming advantage in the form of administrative, judicial, and financial resources. It should be taken into account that modern technologies – not only the combination of outsourcing with the use of public computer networks (Public Data Networks, PDN) to commit a crime, or the possibility of direct remote control of a drone, but also the possibility of using a drone with built-in AI deliberately trained to strike a target – are merely the tip of the iceberg that the “Titanic” of the established security system will collide with. Further technological development, in particular decentralized AI using Web 3.0 / Web3, will make it possible to use AI as the executor of a deceased person’s will, while transferring to the AI the necessary financial resources in cryptocurrency (including programming the AI to further criminal acquisition of funds for its activities), combined with the ability to use fab labs or to have the AI itself hire contractors, creates for the targets of an attack aimed by such an AI a situation of the inevitability of retribution. At the same time, these capabilities can be extrapolated to any life situations – for example, those involving deprivation of liberty, such as in connection with the abduction of any person following the example of the abduction of N. Maduro, or situations such as bankruptcy resulting from the bad-faith actions of counterparties. At the same time, the risk of retribution in the process of defending violated rights affects both rank-and-file executors – such as police officers and judges – and the real masters of the country in the form of the public and non-public elite. The latter situation – the threat to the lives of the elite – already appears to be a real problem requiring a solution. After all, it would be extremely painful for the ruling strata of countries that have fought wars and then reconciled – the main beneficiaries of the past war – to answer to the victims for crimes committed both during the war and during mobilization, even if the terms of peace provide for full amnesty. An absolutely unfamiliar sense of danger will also emerge among the ruling strata governing states that ignite wars and create crises, since they now find themselves in a vulnerable position. This situation is further aggravated by the fact that information – both factual and conspiracy theories – is now widely accessible and can serve as grounds for attacks on representatives of well-known families, both by informed individuals and by mentally ill people. Would the issue of depriving Denmark of Greenland even be on the agenda now if, during the 2024 assassination attempts on D. Trump (AP News, 2025; Reuters, 2025), terrorists had used not firearms but a group of fiber-optic drones with centralized AI trained to recognize its target? The third side of the coin will be the need to minimize offenses in society and to introduce mechanisms of genuine democracy and accountability of the authorities for the results of their activities, when the overwhelming majority of the population is involved in decision-making – from ensuring the functioning of a city district to the election of sheriffs, judges, prosecutors, and all the way to voting on draft laws as well as federal elections (see the experience of Switzerland). This system will make it possible to reduce the number of legal violations by representatives of the ruling strata and to hold them accountable for both past and ongoing crimes without the need for extrajudicial reprisals by private individuals. Concluding the enumeration of the main aspects of changes in public life caused by the development of private combat robotics, let us also consider the fourth side of the same coin. All the technologies and capabilities discussed can be implemented by a wide range of individuals with disturbed psyches, for example religious fanatics, as well as by criminal elements, for whom new technologies present the broadest opportunities for blackmail, robberies, and extortion. And it is precisely against such individuals that it will be necessary to create a security system of a new quality – one that does not yet exist – a security system costing hundreds of billions of euros for each country deploying it, ensuring comprehensive protection of society from new types of threats. Of course, it may seem that the development of such a security system is possible without social modernization of relations in society and without the introduction of mechanisms of real democracy. It may seem that the implementation of a police state based on a digital concentration camp is more preferable. Perhaps – but this would require conducting an experiment, for example following the model of Pakistan or the DPRK, where the ruling military or party elite lives isolated from the main part of the population. In doing so, the ruling stratum would have to survive under new conditions of total war with its own population, from whom, for the sake of “security,” absolutely all remaining freedoms would be taken away, following the example of the DPRK. The application of AI that can operate in our world after the death of the person for whom the AI serves as executor proves that the empirical rule “you can’t take your money with you” is gradually losing its meaning: AI or artificial consciousness (AC) makes it possible to practically and almost inevitably implement the will of either the deceased or, say, a person who has been imprisoned or kidnapped, as well as someone who has found themselves in other situations that limit their ability to act. In this regard, the next customer ordering the next kidnapping of N. Maduro will think very hard about whether it is worth dying from retaliatory actions by AI, or from the actions of an actor who has decided that the triggering event for the AI’s predefined action cycle has occurred. At the same time, an actor in the form of decentralized AI cannot be intimidated, bought, or destroyed. In effect, new technologies put at the disposal of private individuals and organizations what previously only states had at their disposal, in particular an analogue of a system like “Perimeter” (RVSN RF index 15E601, known in journalism as “Dead Hand”) (Stilwell, 2022). Of course, the AI (or AC) systems discussed above – first and foremost decentralized AI, designed so that they cannot be influenced or have their operating order changed either by shutdown or by blackmail involving the risk of shutdown – may also inherently carry socially constructive tasks. Already now, AI systems can function as independent and autonomous executors, even though they still contain certain built-in technological limitations. Even this, however, already makes it possible to use such systems effectively both as operational AI assistants and as systems for auditing human decisions for compliance with specified goals and/or means (Gudkov, 2020; Cowger, 2023; Li, 2024; Bell, 2025; Brennan, 2025; Brown, 2025). Here and throughout, wherever AI is discussed, the possibility of using an IS is also implied – one that differs from AI by the presence of a software equivalent of will. The rate at which AI systems operating in the PDN evolve into IS systems also operating in the PDN is not considered here, nor is the time it takes for laboratory IS systems to enter the PDN. And, as practice shows, innovations are primarily directed toward the sphere of committing crimes – for example, the elimination of undesirable individuals – an activity engaged in both by independent criminals, such as roaming bandits, and by the intelligence services and ministries of defense of stationary bandits – states. In this regard, although the fully robotic technologies discussed above, which do not involve human intervention in their operation from the moment of launch, can also be used for constructive activities, their priority emergence in cri