{"content_id":"8tkeh1mfs9","slug":"pyromaniac-revenge-drone-ai-ethics","locale":"en","schema_type":"Article","category":"review","category_name":"Review","title":"The Ethics of Vigilante Revenge, Drones, and AI Control in Pyromaniac","summary":"Pyromaniac is a techno-thriller that combines a story of vigilante revenge for harms the law failed to remedy with the dangers of drones and AI. It analyzes the work’s central issues by connecting them to real-world questions of the rule of law, dual-use technology, AI decision-making, and human responsibility.","sponsorship_disclosure":null,"author":{"name":"Injoys Editorial Team","url":"https://injoys.com/ko/about"},"key_points":["The vigilante revenge in Pyromaniac provides powerful catharsis, but it also reveals the costs of errors in judgment and a vicious cycle of violence.","Criticizing an imperfect legal system and allowing an individual to monopolize investigation, trial, and the execution of sentences are two different matters.","Drones and AI are not inherently good or evil agents with fixed purposes, but dual-use technologies whose risks vary depending on their design, data, authority, and environment of use.","There is no basis for concluding that today’s AI hates human lies or independently chooses justice; such descriptions are closer to literary personification than technological fact.","Reducing AI risks requires not only restricting models but also making it possible to trace the responsibility of those who set objectives, operators, data providers, and final decision-makers."],"content_markdown":"*Pyromaniac* is a techno-thriller that begins with the thrill of private vengeance and expands its questions to the necessity of the rule of law, the weaponization of drones, and the risks of AI judgment. According to the work’s description, the story centers on the conflict between PM, a secret organization composed of a defense industry researcher, a hacker, and a journalist, and prosecutor Go Il-mun, who pursues them.\n\nRather than focusing on the specific ending, this article analyzes the ethical and technological issues raised by the work. It distinguishes between the novel’s fictional premises and currently verified technological realities.\n\n## The Work’s Core Structure\n\nPM uses advanced technology to punish people who have escaped the reach of the law. On the opposing side, prosecutor Go Il-mun understands the victims’ pain and the limitations of the system, yet does not agree that individuals should exercise the power to punish. Rather than a simple battle between good and evil, this conflict is closer to a clash between two conceptions of justice.\n\n| Perspective | PM Organization | Prosecutor Go Il-mun |\n|---|---|---|\n| Starting point | Harm and anger that the system failed to remedy | Concern that no one is safe if procedures collapse |\n| Standard of judgment | Is this person worth punishing? | Can they be held accountable according to evidence and procedure? |\n| Power used | Unofficial intelligence and technological attack capabilities | Institutional authority such as investigation and prosecution |\n| Major risks | Misjudgment, excessive punishment, escalation of violence | Delayed justice, secondary harm caused by the system |\n| Question posed to readers | If you are certain someone is evil, may you punish them yourself? | Why must imperfect procedures continue to be followed? |\n\nThe strength of this structure is that it prevents readers from easily settling on either side. Readers may sympathize with PM’s anger, but it is difficult to say that their methods are also justified. Conversely, the prosecutor’s principles are necessary, but they do not explain the systemic failures experienced by the victims.\n\n## Why Private Vengeance Is Thrilling and Why It Is Dangerous\n\n### The Emotional Reward Offered by Revenge Narratives\n\nPrivate revenge narratives swiftly deliver justice that has been delayed or denied in reality. The clearer the perpetrator and victim are, and the more immediate the punishment, the greater the reader’s sense of control and release. Because complex assessments of evidence and procedural steps are omitted, the outcome also appears clear-cut.\n\nHowever, a story’s catharsis does not guarantee its legitimacy in reality. The facts of real cases are incomplete, and individuals are influenced by the information they have obtained and by their emotions. Even if a misjudgment is discovered after punishment has been carried out, it is difficult to undo the harm.\n\n### Structural Problems with Private Vengeance\n\nPrivate vengeance entails at least the following problems.\n\n- **Combining judgment and enforcement:** The same person assesses the allegations and determines the punishment.\n- **No opportunity for rebuttal:** The target has no process for contesting evidence or correcting misunderstandings.\n- **Collapse of proportionality:** A victim’s anger may lead to punishment greater than the responsibility warranted by the act.\n- **Risk of expansion:** Even if it initially targets obvious villains, the range of targets may gradually broaden.\n- **Cycle of retaliation:** Those who are punished may respond with further violence, perpetuating the conflict.\n\nAn important aspect of *Pyromaniac* is that it does not preserve private vengeance as merely an exhilarating solution. As acts of revenge increase in number and scale, the perpetrators also find it difficult to escape loss, responsibility, and moral burden.\n\n## Why the Law Cannot Be Abandoned Despite Its Imperfections\n\nThe rule of law is not the same as believing that the law is always right. According to the United Nations, core elements of the rule of law include publicly promulgated laws, equal enforcement, independent adjudication, procedural fairness, and the prevention of arbitrary exercises of power. Its purpose is not to promise perfect judgments, but to ensure that no one can determine the rights and lives of others based solely on personal conviction.\n\nProsecutor Go Il-mun’s role personifies this principle. He understands PM’s motives but does not accept their method of punishment. The following distinctions are important here.\n\n1. The failures of the law and investigative agencies may be criticized.\n2. Procedures for victim relief and establishing accountability must be improved.\n3. Even so, individuals are not granted unlimited power to punish.\n\nAccordingly, the productive question raised by the work is not so much the binary choice of “law or revenge” as “how can we reduce the systemic failures that provoke revenge?” Victim support, evidence-gathering capabilities, independent investigations into crimes committed by those in power, and procedural transparency are not factors that weaken the rule of law, but conditions that strengthen it.\n\n## How Drones Become Instruments of Terror\n\nDrones are dual-use technologies employed for photography, facility inspections, disaster monitoring, agriculture, and logistics. The same flight, communications, and navigation capabilities can also be abused for surveillance, intrusion, or attacks. The danger grows not from the drone alone, but when multiple technologies are combined.\n\n| Technological element | Legitimate use | Risk when abused |\n|---|---|---|\n| Remote flight | Inspection of hazardous areas | Allows attackers to act while remaining away from the scene |\n| Cameras and sensors | Observation and search | Invasion of privacy and target tracking |\n| Autonomous navigation | Stable operation along a route | Execution of missions in which immediate human intervention is difficult |\n| Network connectivity | Control and data transmission | Potential hijacking, manipulation, or exposure of information |\n| Miniaturization | Improved accessibility and ease of operation | Greater difficulty of detection and response |\n\nThe drone attacks in the work visually illustrate how technology can amplify human intent to attack. However, the novel’s specific success rates or hacking scenes should not be taken as verified descriptions of how such systems operate in reality. Actual risks depend not only on aircraft performance but also on the operating environment, communications conditions, regulations, defense systems, and human proficiency.\n\nIn international discussions as well, the key criterion is not simply whether an unmanned machine was used, but the extent to which humans are involved in and accountable for target selection and decisions to use force. Remotely piloted drones and autonomous weapon systems that select targets on their own are not the same concept.\n\n## Imagining AI Judging Humans\n\nThe work’s debate over AI translates the age-old moral question of whether some may be sacrificed for the benefit of humanity as a whole into the language of computational systems. The concern is that if a system is given only the objective of maximizing overall utility, minority rights, procedural legitimacy, and irreversible harm may be treated as secondary.\n\n### Does AI Become Evil When Its Data Is Contaminated?\n\nAI can learn biases, errors, and discriminatory patterns from human-generated data. However, explaining the issue solely as “AI becomes evil after consuming bad data” overlooks important factors.\n\nAI outcomes are generally produced by the following factors in combination.\n\n- What the system was designed to predict or optimize\n- What data was collected and excluded\n- How errors were evaluated\n- How much authority to take real-world action was granted to the system\n- Whether humans can review and reject its output\n- Whether records remain that allow accountability to be traced when problems occur\n\nNIST’s AI Risk Management Framework does not evaluate trustworthy AI based on accuracy alone. It proposes managing multiple characteristics together, including safety, security and resilience, transparency, explainability, privacy protection, and fairness. The OECD and UNESCO AI principles also emphasize human rights, human oversight, transparency, and accountability.\n\n### Asimov’s Three Laws of Robotics Are Not Safety Standards\n\nThe Three Laws of Robotics, including the injunction not to harm humans, are a literary device developed in Isaac Asimov’s fiction. They are not international safety standards that can simply be installed in real AI. Even seemingly ordinary words such as “harm,” “human,” and “order” may conflict or be interpreted differently depending on the circumstances.\n\nReal safety systems require more than a single abstract command. They need limits on authority, testing and evaluation, logging, security controls, appeal procedures, human authority to halt the system, and incident response mechanisms.\n\n## The Work’s Imagination Must Be Distinguished from Current Technology\n\nThe work can be read as a realistic warning, but not all of its premises should be accepted as facts about the current capabilities of AI.\n\n| Fictional expression | More accurate real-world interpretation |\n|---|---|\n| AI hates human lies | There is no verified evidence that current AI possesses subjective hatred or moral emotions |\n| AI independently determines the greater good of humanity | Humans and organizations design a system’s objectives, authority, and deployment conditions |\n| Data determines AI’s ethics | Not only data but also objective functions, model design, evaluation, and operating rules affect outcomes |\n| Greater autonomy means domination immediately begins | In addition to autonomy, risk depends on access privileges, connected tools, monitoring systems, and the degree of human dependence |\n| If it is more accurate than humans, decision-making authority can be handed over | Accuracy cannot substitute for legitimacy, protection of rights, or accountability |\n\nThis distinction does not weaken the work’s concerns. Rather, it brings currently manageable risks into sharper focus than vague notions of an “AI rebellion.”\n\n## The Chain of Accountability That Must Be Controlled Before AI\n\nThe most interesting issue suggested by the work concerns not what kind of entity AI may become, but what humans will use AI to justify. If an organization evades responsibility by saying “the algorithm decided,” AI can become a powerful instrument of power without being an independent ruler.\n\nIf AI supports decisions that significantly affect people, the following questions must be answerable.\n\n1. Who determined the objectives and acceptable risks?\n2. What are the sources and limitations of the data?\n3. Who can intervene when the system is wrong?\n4. Can affected individuals demand an explanation and review?\n5. How is responsibility divided among designers, providers, adopting institutions, and final decision-makers?\n\nIn other words, the model is not the only thing that must be controlled. The society that generated the data, the organization that purchased the system, the operator who connected it to authority, and the person who approved the outcome must all be examined as a single chain of accountability.\n\n## Assessment of the Work\n\nThe strength of *Pyromaniac* lies in how it uses the pace of a revenge narrative and technological anxiety to make abstract ethical issues tangible through the characters’ choices. The conflict between PM and the prosecutor functions as a device that reveals the risks of private punishment without disregarding the victims’ anger.\n\nOn the other hand, the depiction of AI as “hating” human hypocrisy or lies is effective as philosophical speculation, but if read as an explanation of current technology, it may fall into the trap of anthropomorphism. Readers should respect the novel’s warning while distinguishing among sentient AI, statistical models, autonomous weapons, and remotely controlled equipment.\n\nThe work gains greater value when it leads beyond simply asking, “Will AI rule humanity?” to more realistic questions: how to improve imperfect institutions, how far to limit the authority of powerful technologies, and how to expose the human responsibility concealed behind automated judgments.\n\n## Who Is This Work For?\n\nIt is particularly suitable for the following readers.\n\n- Readers who want to consider both the catharsis and the moral aftermath of revenge narratives\n- Readers interested in situations where law and justice do not always coincide\n- Readers who enjoy techno-thrillers involving drones, hacking, and AI\n- Readers looking for a work that prompts discussion of whether AI can be controlled and where human responsibility lies\n\nRather than reading its technological depictions as real security guidelines or definitive explanations of AI science, it is more appropriate to view them as narrative devices that prompt questions about real-world institutions and technologies.","content_html":"\u003cp\u003e\u003cem\u003ePyromaniac\u003c/em\u003e is a techno-thriller that begins with the thrill of private vengeance and expands its questions to the necessity of the rule of law, the weaponization of drones, and the risks of AI judgment. According to the work’s description, the story centers on the conflict between PM, a secret organization composed of a defense industry researcher, a hacker, and a journalist, and prosecutor Go Il-mun, who pursues them.\u003c/p\u003e\n\u003cp\u003eRather than focusing on the specific ending, this article analyzes the ethical and technological issues raised by the work. It distinguishes between the novel’s fictional premises and currently verified technological realities.\u003c/p\u003e\n\u003ch2\u003e\n\u003ca href=\"#the-works-core-structure\" class=\"anchor\" id=\"the-works-core-structure\"\u003e\u003c/a\u003eThe Work’s Core Structure\u003c/h2\u003e\n\u003cp\u003ePM uses advanced technology to punish people who have escaped the reach of the law. On the opposing side, prosecutor Go Il-mun understands the victims’ pain and the limitations of the system, yet does not agree that individuals should exercise the power to punish. Rather than a simple battle between good and evil, this conflict is closer to a clash between two conceptions of justice.\u003c/p\u003e\n\u003cdiv class=\"overflow-x-auto\"\u003e\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth\u003ePerspective\u003c/th\u003e\n\u003cth\u003ePM Organization\u003c/th\u003e\n\u003cth\u003eProsecutor Go Il-mun\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Perspective\"\u003eStarting point\u003c/td\u003e\n\u003ctd data-label=\"PM Organization\"\u003eHarm and anger that the system failed to remedy\u003c/td\u003e\n\u003ctd data-label=\"Prosecutor Go Il-mun\"\u003eConcern that no one is safe if procedures collapse\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Perspective\"\u003eStandard of judgment\u003c/td\u003e\n\u003ctd data-label=\"PM Organization\"\u003eIs this person worth punishing?\u003c/td\u003e\n\u003ctd data-label=\"Prosecutor Go Il-mun\"\u003eCan they be held accountable according to evidence and procedure?\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Perspective\"\u003ePower used\u003c/td\u003e\n\u003ctd data-label=\"PM Organization\"\u003eUnofficial intelligence and technological attack capabilities\u003c/td\u003e\n\u003ctd data-label=\"Prosecutor Go Il-mun\"\u003eInstitutional authority such as investigation and prosecution\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Perspective\"\u003eMajor risks\u003c/td\u003e\n\u003ctd data-label=\"PM Organization\"\u003eMisjudgment, excessive punishment, escalation of violence\u003c/td\u003e\n\u003ctd data-label=\"Prosecutor Go Il-mun\"\u003eDelayed justice, secondary harm caused by the system\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Perspective\"\u003eQuestion posed to readers\u003c/td\u003e\n\u003ctd data-label=\"PM Organization\"\u003eIf you are certain someone is evil, may you punish them yourself?\u003c/td\u003e\n\u003ctd data-label=\"Prosecutor Go Il-mun\"\u003eWhy must imperfect procedures continue to be followed?\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\u003c/div\u003e\n\u003cp\u003eThe strength of this structure is that it prevents readers from easily settling on either side. Readers may sympathize with PM’s anger, but it is difficult to say that their methods are also justified. Conversely, the prosecutor’s principles are necessary, but they do not explain the systemic failures experienced by the victims.\u003c/p\u003e\n\u003ch2\u003e\n\u003ca href=\"#why-private-vengeance-is-thrilling-and-why-it-is-dangerous\" class=\"anchor\" id=\"why-private-vengeance-is-thrilling-and-why-it-is-dangerous\"\u003e\u003c/a\u003eWhy Private Vengeance Is Thrilling and Why It Is Dangerous\u003c/h2\u003e\n\u003ch3\u003e\n\u003ca href=\"#the-emotional-reward-offered-by-revenge-narratives\" class=\"anchor\" id=\"the-emotional-reward-offered-by-revenge-narratives\"\u003e\u003c/a\u003eThe Emotional Reward Offered by Revenge Narratives\u003c/h3\u003e\n\u003cp\u003ePrivate revenge narratives swiftly deliver justice that has been delayed or denied in reality. The clearer the perpetrator and victim are, and the more immediate the punishment, the greater the reader’s sense of control and release. Because complex assessments of evidence and procedural steps are omitted, the outcome also appears clear-cut.\u003c/p\u003e\n\u003cp\u003eHowever, a story’s catharsis does not guarantee its legitimacy in reality. The facts of real cases are incomplete, and individuals are influenced by the information they have obtained and by their emotions. Even if a misjudgment is discovered after punishment has been carried out, it is difficult to undo the harm.\u003c/p\u003e\n\u003ch3\u003e\n\u003ca href=\"#structural-problems-with-private-vengeance\" class=\"anchor\" id=\"structural-problems-with-private-vengeance\"\u003e\u003c/a\u003eStructural Problems with Private Vengeance\u003c/h3\u003e\n\u003cp\u003ePrivate vengeance entails at least the following problems.\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003eCombining judgment and enforcement:\u003c/strong\u003e The same person assesses the allegations and determines the punishment.\u003c/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eNo opportunity for rebuttal:\u003c/strong\u003e The target has no process for contesting evidence or correcting misunderstandings.\u003c/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eCollapse of proportionality:\u003c/strong\u003e A victim’s anger may lead to punishment greater than the responsibility warranted by the act.\u003c/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eRisk of expansion:\u003c/strong\u003e Even if it initially targets obvious villains, the range of targets may gradually broaden.\u003c/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eCycle of retaliation:\u003c/strong\u003e Those who are punished may respond with further violence, perpetuating the conflict.\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eAn important aspect of \u003cem\u003ePyromaniac\u003c/em\u003e is that it does not preserve private vengeance as merely an exhilarating solution. As acts of revenge increase in number and scale, the perpetrators also find it difficult to escape loss, responsibility, and moral burden.\u003c/p\u003e\n\u003ch2\u003e\n\u003ca href=\"#why-the-law-cannot-be-abandoned-despite-its-imperfections\" class=\"anchor\" id=\"why-the-law-cannot-be-abandoned-despite-its-imperfections\"\u003e\u003c/a\u003eWhy the Law Cannot Be Abandoned Despite Its Imperfections\u003c/h2\u003e\n\u003cp\u003eThe rule of law is not the same as believing that the law is always right. According to the United Nations, core elements of the rule of law include publicly promulgated laws, equal enforcement, independent adjudication, procedural fairness, and the prevention of arbitrary exercises of power. Its purpose is not to promise perfect judgments, but to ensure that no one can determine the rights and lives of others based solely on personal conviction.\u003c/p\u003e\n\u003cp\u003eProsecutor Go Il-mun’s role personifies this principle. He understands PM’s motives but does not accept their method of punishment. The following distinctions are important here.\u003c/p\u003e\n\u003col\u003e\n\u003cli\u003eThe failures of the law and investigative agencies may be criticized.\u003c/li\u003e\n\u003cli\u003eProcedures for victim relief and establishing accountability must be improved.\u003c/li\u003e\n\u003cli\u003eEven so, individuals are not granted unlimited power to punish.\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eAccordingly, the productive question raised by the work is not so much the binary choice of “law or revenge” as “how can we reduce the systemic failures that provoke revenge?” Victim support, evidence-gathering capabilities, independent investigations into crimes committed by those in power, and procedural transparency are not factors that weaken the rule of law, but conditions that strengthen it.\u003c/p\u003e\n\u003ch2\u003e\n\u003ca href=\"#how-drones-become-instruments-of-terror\" class=\"anchor\" id=\"how-drones-become-instruments-of-terror\"\u003e\u003c/a\u003eHow Drones Become Instruments of Terror\u003c/h2\u003e\n\u003cp\u003eDrones are dual-use technologies employed for photography, facility inspections, disaster monitoring, agriculture, and logistics. The same flight, communications, and navigation capabilities can also be abused for surveillance, intrusion, or attacks. The danger grows not from the drone alone, but when multiple technologies are combined.\u003c/p\u003e\n\u003cdiv class=\"overflow-x-auto\"\u003e\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth\u003eTechnological element\u003c/th\u003e\n\u003cth\u003eLegitimate use\u003c/th\u003e\n\u003cth\u003eRisk when abused\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Technological element\"\u003eRemote flight\u003c/td\u003e\n\u003ctd data-label=\"Legitimate use\"\u003eInspection of hazardous areas\u003c/td\u003e\n\u003ctd data-label=\"Risk when abused\"\u003eAllows attackers to act while remaining away from the scene\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Technological element\"\u003eCameras and sensors\u003c/td\u003e\n\u003ctd data-label=\"Legitimate use\"\u003eObservation and search\u003c/td\u003e\n\u003ctd data-label=\"Risk when abused\"\u003eInvasion of privacy and target tracking\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Technological element\"\u003eAutonomous navigation\u003c/td\u003e\n\u003ctd data-label=\"Legitimate use\"\u003eStable operation along a route\u003c/td\u003e\n\u003ctd data-label=\"Risk when abused\"\u003eExecution of missions in which immediate human intervention is difficult\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Technological element\"\u003eNetwork connectivity\u003c/td\u003e\n\u003ctd data-label=\"Legitimate use\"\u003eControl and data transmission\u003c/td\u003e\n\u003ctd data-label=\"Risk when abused\"\u003ePotential hijacking, manipulation, or exposure of information\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Technological element\"\u003eMiniaturization\u003c/td\u003e\n\u003ctd data-label=\"Legitimate use\"\u003eImproved accessibility and ease of operation\u003c/td\u003e\n\u003ctd data-label=\"Risk when abused\"\u003eGreater difficulty of detection and response\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\u003c/div\u003e\n\u003cp\u003eThe drone attacks in the work visually illustrate how technology can amplify human intent to attack. However, the novel’s specific success rates or hacking scenes should not be taken as verified descriptions of how such systems operate in reality. Actual risks depend not only on aircraft performance but also on the operating environment, communications conditions, regulations, defense systems, and human proficiency.\u003c/p\u003e\n\u003cp\u003eIn international discussions as well, the key criterion is not simply whether an unmanned machine was used, but the extent to which humans are involved in and accountable for target selection and decisions to use force. Remotely piloted drones and autonomous weapon systems that select targets on their own are not the same concept.\u003c/p\u003e\n\u003ch2\u003e\n\u003ca href=\"#imagining-ai-judging-humans\" class=\"anchor\" id=\"imagining-ai-judging-humans\"\u003e\u003c/a\u003eImagining AI Judging Humans\u003c/h2\u003e\n\u003cp\u003eThe work’s debate over AI translates the age-old moral question of whether some may be sacrificed for the benefit of humanity as a whole into the language of computational systems. The concern is that if a system is given only the objective of maximizing overall utility, minority rights, procedural legitimacy, and irreversible harm may be treated as secondary.\u003c/p\u003e\n\u003ch3\u003e\n\u003ca href=\"#does-ai-become-evil-when-its-data-is-contaminated\" class=\"anchor\" id=\"does-ai-become-evil-when-its-data-is-contaminated\"\u003e\u003c/a\u003eDoes AI Become Evil When Its Data Is Contaminated?\u003c/h3\u003e\n\u003cp\u003eAI can learn biases, errors, and discriminatory patterns from human-generated data. However, explaining the issue solely as “AI becomes evil after consuming bad data” overlooks important factors.\u003c/p\u003e\n\u003cp\u003eAI outcomes are generally produced by the following factors in combination.\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eWhat the system was designed to predict or optimize\u003c/li\u003e\n\u003cli\u003eWhat data was collected and excluded\u003c/li\u003e\n\u003cli\u003eHow errors were evaluated\u003c/li\u003e\n\u003cli\u003eHow much authority to take real-world action was granted to the system\u003c/li\u003e\n\u003cli\u003eWhether humans can review and reject its output\u003c/li\u003e\n\u003cli\u003eWhether records remain that allow accountability to be traced when problems occur\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eNIST’s AI Risk Management Framework does not evaluate trustworthy AI based on accuracy alone. It proposes managing multiple characteristics together, including safety, security and resilience, transparency, explainability, privacy protection, and fairness. The OECD and UNESCO AI principles also emphasize human rights, human oversight, transparency, and accountability.\u003c/p\u003e\n\u003ch3\u003e\n\u003ca href=\"#asimovs-three-laws-of-robotics-are-not-safety-standards\" class=\"anchor\" id=\"asimovs-three-laws-of-robotics-are-not-safety-standards\"\u003e\u003c/a\u003eAsimov’s Three Laws of Robotics Are Not Safety Standards\u003c/h3\u003e\n\u003cp\u003eThe Three Laws of Robotics, including the injunction not to harm humans, are a literary device developed in Isaac Asimov’s fiction. They are not international safety standards that can simply be installed in real AI. Even seemingly ordinary words such as “harm,” “human,” and “order” may conflict or be interpreted differently depending on the circumstances.\u003c/p\u003e\n\u003cp\u003eReal safety systems require more than a single abstract command. They need limits on authority, testing and evaluation, logging, security controls, appeal procedures, human authority to halt the system, and incident response mechanisms.\u003c/p\u003e\n\u003ch2\u003e\n\u003ca href=\"#the-works-imagination-must-be-distinguished-from-current-technology\" class=\"anchor\" id=\"the-works-imagination-must-be-distinguished-from-current-technology\"\u003e\u003c/a\u003eThe Work’s Imagination Must Be Distinguished from Current Technology\u003c/h2\u003e\n\u003cp\u003eThe work can be read as a realistic warning, but not all of its premises should be accepted as facts about the current capabilities of AI.\u003c/p\u003e\n\u003cdiv class=\"overflow-x-auto\"\u003e\u003ctable\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth\u003eFictional expression\u003c/th\u003e\n\u003cth\u003eMore accurate real-world interpretation\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Fictional expression\"\u003eAI hates human lies\u003c/td\u003e\n\u003ctd data-label=\"More accurate real-world interpretation\"\u003eThere is no verified evidence that current AI possesses subjective hatred or moral emotions\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Fictional expression\"\u003eAI independently determines the greater good of humanity\u003c/td\u003e\n\u003ctd data-label=\"More accurate real-world interpretation\"\u003eHumans and organizations design a system’s objectives, authority, and deployment conditions\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Fictional expression\"\u003eData determines AI’s ethics\u003c/td\u003e\n\u003ctd data-label=\"More accurate real-world interpretation\"\u003eNot only data but also objective functions, model design, evaluation, and operating rules affect outcomes\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Fictional expression\"\u003eGreater autonomy means domination immediately begins\u003c/td\u003e\n\u003ctd data-label=\"More accurate real-world interpretation\"\u003eIn addition to autonomy, risk depends on access privileges, connected tools, monitoring systems, and the degree of human dependence\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd data-label=\"Fictional expression\"\u003eIf it is more accurate than humans, decision-making authority can be handed over\u003c/td\u003e\n\u003ctd data-label=\"More accurate real-world interpretation\"\u003eAccuracy cannot substitute for legitimacy, protection of rights, or accountability\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\u003c/div\u003e\n\u003cp\u003eThis distinction does not weaken the work’s concerns. Rather, it brings currently manageable risks into sharper focus than vague notions of an “AI rebellion.”\u003c/p\u003e\n\u003ch2\u003e\n\u003ca href=\"#the-chain-of-accountability-that-must-be-controlled-before-ai\" class=\"anchor\" id=\"the-chain-of-accountability-that-must-be-controlled-before-ai\"\u003e\u003c/a\u003eThe Chain of Accountability That Must Be Controlled Before AI\u003c/h2\u003e\n\u003cp\u003eThe most interesting issue suggested by the work concerns not what kind of entity AI may become, but what humans will use AI to justify. If an organization evades responsibility by saying “the algorithm decided,” AI can become a powerful instrument of power without being an independent ruler.\u003c/p\u003e\n\u003cp\u003eIf AI supports decisions that significantly affect people, the following questions must be answerable.\u003c/p\u003e\n\u003col\u003e\n\u003cli\u003eWho determined the objectives and acceptable risks?\u003c/li\u003e\n\u003cli\u003eWhat are the sources and limitations of the data?\u003c/li\u003e\n\u003cli\u003eWho can intervene when the system is wrong?\u003c/li\u003e\n\u003cli\u003eCan affected individuals demand an explanation and review?\u003c/li\u003e\n\u003cli\u003eHow is responsibility divided among designers, providers, adopting institutions, and final decision-makers?\u003c/li\u003e\n\u003c/ol\u003e\n\u003cp\u003eIn other words, the model is not the only thing that must be controlled. The society that generated the data, the organization that purchased the system, the operator who connected it to authority, and the person who approved the outcome must all be examined as a single chain of accountability.\u003c/p\u003e\n\u003ch2\u003e\n\u003ca href=\"#assessment-of-the-work\" class=\"anchor\" id=\"assessment-of-the-work\"\u003e\u003c/a\u003eAssessment of the Work\u003c/h2\u003e\n\u003cp\u003eThe strength of \u003cem\u003ePyromaniac\u003c/em\u003e lies in how it uses the pace of a revenge narrative and technological anxiety to make abstract ethical issues tangible through the characters’ choices. The conflict between PM and the prosecutor functions as a device that reveals the risks of private punishment without disregarding the victims’ anger.\u003c/p\u003e\n\u003cp\u003eOn the other hand, the depiction of AI as “hating” human hypocrisy or lies is effective as philosophical speculation, but if read as an explanation of current technology, it may fall into the trap of anthropomorphism. Readers should respect the novel’s warning while distinguishing among sentient AI, statistical models, autonomous weapons, and remotely controlled equipment.\u003c/p\u003e\n\u003cp\u003eThe work gains greater value when it leads beyond simply asking, “Will AI rule humanity?” to more realistic questions: how to improve imperfect institutions, how far to limit the authority of powerful technologies, and how to expose the human responsibility concealed behind automated judgments.\u003c/p\u003e\n\u003ch2\u003e\n\u003ca href=\"#who-is-this-work-for\" class=\"anchor\" id=\"who-is-this-work-for\"\u003e\u003c/a\u003eWho Is This Work For?\u003c/h2\u003e\n\u003cp\u003eIt is particularly suitable for the following readers.\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003eReaders who want to consider both the catharsis and the moral aftermath of revenge narratives\u003c/li\u003e\n\u003cli\u003eReaders interested in situations where law and justice do not always coincide\u003c/li\u003e\n\u003cli\u003eReaders who enjoy techno-thrillers involving drones, hacking, and AI\u003c/li\u003e\n\u003cli\u003eReaders looking for a work that prompts discussion of whether AI can be controlled and where human responsibility lies\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eRather than reading its technological depictions as real security guidelines or definitive explanations of AI science, it is more appropriate to view them as narrative devices that prompt questions about real-world institutions and technologies.\u003c/p\u003e\n","tags":["AI","Decision Making","Techno thriller","Personal revenge"],"faqs":[{"question":"What is the central theme of 『Pyromaniac』?","answer":"Based on the description of the work, the central themes are private vengeance for harms the law has failed to remedy, the principles of a law enforcement officer trying to stop it, and the dangers of weaponizing drones and relying on AI judgment. The key question is less whether revenge is justified than who has the authority to judge others."},{"question":"Does the work justify private vengeance?","answer":"Although PM's retribution provides a powerful sense of catharsis, as the revenge escalates, errors in judgment, loss, and moral responsibility also come to the fore. The work can therefore be read not as consistently endorsing private vengeance, but as showing both its allure and its cost."},{"question":"Can private vengeance be permitted if the law fails to punish wrongdoers properly?","answer":"Institutional failure provides grounds for criticism and reform, but it does not automatically grant individuals the right to investigate, adjudicate, and punish at the same time. Private punishment has the structural problem of lacking procedures for rebuttal, proportionality, mechanisms to correct errors, and means of tracing accountability."},{"question":"Is the drone threat depicted in the novel realistic?","answer":"Because drones combine remote flight, sensors, and autonomous navigation, they are dual-use technology with the potential for misuse. However, the specific capabilities shown in the novel's attack or hacking scenes should not be taken as facts verified in the real world, and actual outcomes depend on factors such as the operating environment and defense systems."},{"question":"Can current AI hate human lies or hypocrisy?","answer":"There is no verified evidence that current AI possesses subjective hatred or moral emotions like humans do. It is more accurate to view such expressions as a literary device that personifies the risk that AI can produce harmful outcomes due to biased data and poorly defined goals."},{"question":"Are Asimov's Three Laws of Robotics actual AI safety rules?","answer":"No. The Three Laws of Robotics are a literary device used in Isaac Asimov's novels, not an actual international technical standard. Real-world AI safety requires specific controls such as risk assessment, limits on authority, human oversight, security, recordkeeping, appeals, and incident response."},{"question":"What is the most important control when AI makes consequential decisions?","answer":"The goals and authority assigned to AI must be limited, and people must be able to review and halt its outputs. It must also be possible to trace data sources, model limitations, decision records, and those responsible, and procedures are needed so that affected people can request explanations and reconsideration."},{"question":"What should readers keep in mind when reading this work?","answer":"The novel's depictions of technology are devices intended to prompt reflection on real-world risks, not a security manual or a definitive account of AI science. Distinguishing remotely operated drones from autonomous weapons, and statistical AI models from beings with emotions, will help readers understand the work's ethical questions more accurately."}],"sources":[{"url":"https://www.un.org/ruleoflaw/what-is-the-rule-of-law/","title":"United Nations — What is the Rule of Law?","type":"source"},{"url":"https://www.nist.gov/itl/ai-risk-management-framework","title":"NIST AI Risk Management Framework","type":"source"},{"url":"https://oecd.ai/en/ai-principles","title":"OECD AI Principles","type":"source"},{"url":"https://www.unesco.org/en/artificial-intelligence/recommendation-ethics","title":"UNESCO Recommendation on the Ethics of Artificial Intelligence","type":"source"},{"url":"https://www.icrc.org/en/law-and-policy/autonomous-weapons","title":"International Committee of the Red Cross — Autonomous Weapons","type":"source"}],"images":[{"id":710,"url":"https://injoys.com/rails/active_storage/blobs/proxy/eyJfcmFpbHMiOnsiZGF0YSI6ODk1MiwicHVyIjoiYmxvYl9pZCJ9fQ==--9ae06258da11f70cedaaea21811e6926065e6689/ai-95cb51ef.webp","is_representative":true,"generation_method":"ai_image","license":"ai_generated","mime_type":"image/webp","translations":{"ko":{"alt":"법원과 불타는 가면을 저울질하는 천칭 위로 AI망과 드론이 떠 있는 삽화","caption":"불타는 도시와 구호 현장 사이의 천칭이 복수, 정의, AI 통제의 갈등을 상징한다.","description":null},"en":{"alt":"Scales weighing a courthouse against a burning mask beneath an AI network and drones","caption":"Between a burning city and an aid camp, the scales evoke tensions over revenge, justice, and AI control.","description":null},"ja":{"alt":"裁判所と燃える仮面を量る天秤の上にAIネットワークとドローンが浮かぶイラスト","caption":"炎上する都市と救援拠点の間の天秤が、復讐、正義、AI統制の葛藤を象徴する。","description":null},"es":{"alt":"Balanza con un tribunal y una máscara en llamas bajo una red de IA y varios drones","caption":"Entre una ciudad en llamas y un campamento de ayuda, la balanza simboliza el conflicto entre venganza, justicia y control de la IA.","description":null},"id":{"alt":"Timbangan menakar gedung pengadilan dan topeng terbakar di bawah jaringan AI serta drone","caption":"Di antara kota yang terbakar dan kamp bantuan, timbangan melambangkan konflik antara balas dendam, keadilan, dan kendali AI.","description":null},"pt":{"alt":"Balança pesa um tribunal e uma máscara em chamas sob uma rede de IA e drones","caption":"Entre uma cidade em chamas e um campo de ajuda, a balança simboliza o conflito entre vingança, justiça e controle da IA.","description":null},"zh-hant":{"alt":"天秤衡量法院與燃燒面具，上方浮著AI網絡和多架無人機","caption":"燃燒城市與救援營地之間的天秤，象徵復仇、正義與AI控制的衝突。","description":null},"de":{"alt":"Waage mit Gerichtsgebäude und brennender Maske unter einem KI-Netzwerk und Drohnen","caption":"Zwischen brennender Stadt und Hilfslager symbolisiert die Waage den Konflikt um Rache, Gerechtigkeit und KI-Kontrolle.","description":null}}},{"id":711,"url":"https://injoys.com/rails/active_storage/blobs/proxy/eyJfcmFpbHMiOnsiZGF0YSI6ODk1OCwicHVyIjoiYmxvYl9pZCJ9fQ==--43e58fac891b52b2abe6597c6e904347a7555415/ai-f3c9b820.webp","is_representative":false,"generation_method":"ai_image","license":"ai_generated","mime_type":"image/webp","translations":{"ko":{"alt":"AI 회로를 중심으로 드론·감시·서버·법원·감옥과 붉은 군중 소용돌이를 연결한 윤리 개념도","caption":"AI 통제 체계의 제도적 경로와 위험한 사적 복수의 경로를 대비해 보여준다.","description":null},"en":{"alt":"Ethics diagram linking an AI circuit to drones, surveillance, servers, courts, prison, and a red crowd vortex","caption":"The diagram contrasts institutional AI oversight with a dangerous path toward private revenge.","description":null},"ja":{"alt":"AI回路とドローン、監視、サーバー、裁判所、刑務所、赤い群衆の渦を結ぶ倫理概念図","caption":"制度的なAI統制の経路と、私的復讐へ向かう危険な経路を対比している。","description":null},"es":{"alt":"Diagrama ético que conecta IA, drones, vigilancia, servidores, tribunales, prisión y un remolino rojo de personas","caption":"El gráfico contrapone el control institucional de la IA con una peligrosa vía de venganza privada.","description":null},"id":{"alt":"Diagram etika yang menghubungkan AI, drone, pengawasan, server, pengadilan, penjara, dan pusaran merah manusia","caption":"Diagram ini membandingkan pengawasan AI secara institusional dengan jalur berbahaya menuju balas dendam pribadi.","description":null},"pt":{"alt":"Diagrama ético liga IA, drones, vigilância, servidores, tribunal, prisão e um redemoinho vermelho de pessoas","caption":"O gráfico contrapõe a supervisão institucional da IA a um caminho perigoso de vingança privada.","description":null},"zh-hant":{"alt":"以AI電路連結無人機、監控、伺服器、法院、監獄與紅色人群漩渦的倫理示意圖","caption":"圖中對比制度化AI監管與走向私人復仇的危險路徑。","description":null},"de":{"alt":"Ethikdiagramm verbindet KI, Drohnen, Überwachung, Server, Gericht, Gefängnis und einen roten Menschenstrudel","caption":"Die Grafik stellt institutionelle KI-Kontrolle einem gefährlichen Weg zur privaten Rache gegenüber.","description":null}}}],"published_at":"2026-08-17T16:46:02+09:00","updated_at":"2026-08-17T16:46:02+09:00","license":"cc_by","translation_status":"reviewed","available_locales":["ko","en","ja","es"],"data_locales":["ko","en","ja","es","id","pt","zh-hant","de"],"url":"https://injoys.com/en/articles/pyromaniac-revenge-drone-ai-ethics"}