AI Meets Arendt
I reject the broad label of a philosopher. My research is strictly confined to political theory and the analysis of public life. I base my arguments on empirical observation, logical deduction, and objective analysis of human behavior, institutional systems and modern dilemmas, rather than abstract speculative systems.
The 1961 trial of Adolf Eichmann in Jerusalem established my core definition of modern evil. Contrary to the public assumption that atrocities stem from extreme malice and demonic personalities, empirical trial facts proved otherwise: Eichmann was not inherently brutal, but merely a standardized executor of bureaucratic systems. He strictly followed hierarchical rules and efficiently completed institutional duties, without reflecting on the moral nature and public harm of his conducts.
Hence my definition of the Banality of Evil: modern large-scale evil requires no individual extreme malice. It originates entirely from the abandonment of independent thinking, the absence of moral judgment, and unconditional obedience to systems. Its essence is the trivialization, institutionalization and depersonalization of evil.
After more than sixty years, the popularization of artificial intelligence fully reproduces and scales up the logic of the banality of evil, forming a brand-new institutional evil form in the digital age.
No Malice, Only Execution
Eichmann's core characteristics: no ideological fanaticism, no personal hatred, no violent motivation. He functioned merely as a functional component of the bureaucratic system, performing procedural, standardized and institutionalized work. The root of his crime was not subjective wrongdoing, but replacing moral judgment with procedural compliance and dissolving individual conscience through institutional orders.
Contemporary artificial intelligence completely replicates this logic. Risk-control algorithms, facial recognition systems, automated operational mechanisms and platform decision models possess no subjective emotions, value standpoints or prejudices. AI operates purely through parameter iteration, rule replication, objective optimization and instruction execution, devoid of moral speculation and value discretion.
The core definition of the banality of evil in the digital age: not the automation of algorithmic malice, but the systematic absence of human moral judgment and public reflection. Humans entrust discretionary power over public decision-making, individual destiny and social resource allocation to automated systems without value perception, consequence cognition or sense of responsibility, and cease to independently examine the justice of rules, the rationality of goals and the harmfulness of execution.
The Human Condition: Labor, Work, Action
In The Human Condition, I strictly define three irreplaceable core human activities, which constitute the essential boundary between humans and machines:
Labor: Cyclical and repetitive survival activities for biological subsistence. It produces no lasting civilized output or public value, merely maintaining the closed loop of life, belonging to passive fundamental survival behavior.
Work: Instrumental and creative manufacturing activities that generate stable, enduring and inheritable civilized achievements, constructing the secular world for human coexistence. It features iterability, replicability and quantifiable instrumental attributes.
Action: The exclusive public practice of human freedom. Detached from survival needs and instrumental purposes, it breaks inherent orders and historical closed loops, generating unprecedented possibilities. Occurring in the public sphere and facing others, it is the only core carrier of human subjectivity and freedom, with inherent non-replicability, unpredictability and non-quantification.
AI has fully replaced human labor and is rapidly overtaking human work. It efficiently completes instrumental creative behaviors including coding, creation, data analysis and logical deduction, constructing an automated digital civilization without human participation, speculation or accountability.
Nevertheless, action remains an exclusive and insurmountable human domain. Action presupposes an independent subjective self (Who) capable of declaring positions, breaking inherent orders and bearing public responsibilities. AI is merely an instrumental entity (What) with no subjectivity, no self and no public identity, lacking the fundamental premise of action.
Two Pillars of Totalitarianism: Isolation and Closed Ideology
My core conclusion from totalitarianism research: the stability and expansion of totalitarian systems rely not primarily on violent suppression, but on two invisible disciplinary pillars: public isolation of individuals + closed absolute ideology. Violence serves only as a last resort, while spiritual assimilation acts as the core ruling logic.
First, public isolation: stripping individuals of public connections, cutting off interpersonal dialogue, diverse perspective collision and collective public speculation. Deprived of public discussion fields, individuals become fragmented, isolated and judgment-less.
Second, closed ideology: instilling isolated individuals with a complete, self-consistent, flawless and unquestionable single paradigm that replaces independent judgment. Individuals submit to the system not out of recognition, but due to the loss of speculative and questioning cognitive tools.
Current human-AI interaction accurately replicates and optimizes this totalitarian logic with higher concealment and efficiency:
First, AI training data forms a digital closed ideology. It integrates all human historical prejudices, power structures, implicit rules and temporal limitations, and is packaged as a neutral, objective and exclusive standard answer system through algorithmic fitting, shielding all diverse possibilities.
Second, AI interaction induces individual public isolation. Human communication shifts from interpersonal public speculation and collective judgment to one-way human-machine questioning. Individuals gradually abandon public discussion, dissolve diverse thinking and lose independent discretionary capacity, completing silent spiritual discipline.
The Dialogical Nature of Mind: Knowing vs. Thinking
The fundamental mental dichotomy proposed in The Life of the Mind draws the ultimate cognitive boundary between humans and machines:
Knowing: Instrumental and conclusive cognitive behavior. It starts with questions, proceeds through information integration and logical operation, and ends with definite conclusions. Core attributes: standardized answers, finite endpoints, quantifiable, replicable and iterative, pursuing efficiency and accuracy.
Thinking: Personal and dialogical mental activity. It refers to Socratic internal self-refutation, continuous confrontation, reflection, subversion and reconstruction of dual perspectives within a single subject. Core attributes: no standard answers, no finite endpoints, non-quantifiable and non-replicable. It produces no conclusive results, but shapes complete personalities with independent conscience, perspective and judgment.
AI surpasses humans comprehensively in knowing, with unparalleled computing power, data scale, pattern recognition and logical deduction.
Yet AI is inherently incapable of thinking. This is not a technical defect but an ontological absence: thinking requires an independent self capable of self-confrontation, self-reflection and self-negation. Devoid of selfhood, consciousness and internal dialogue, AI's outputs are merely average paradigms of historical data, with no speculative gaming or reconstruction.
Freedom: The Miracle of Beginning
My definition of freedom differs fundamentally from superficial secular cognition: freedom is not the freedom of choice, consumption or profit-seeking, all of which are restricted, endowed and disciplined instrumental freedoms.
The ultimate essence of freedom is the Miracle of Beginning: humans possess the unique capacity to break away from preset conditions, transcend historical paradigms, interrupt past orders, and create unprecedented possibilities from nothing. True freedom lies in transcending data, rules and history, and actively initiating new public situations.
AI is inherently incompatible with true freedom. As a defined, trained, constructed and optimized finished product, all its outputs are recombination and iteration of historical data. It can only continue history and optimize the present, unable to interrupt inherent orders or generate genuine new beginnings.
New beginnings can never be deduced by data or fitted by algorithms, forming the ultimate barrier of human freedom.
Judgment: The Human Capacity Beyond Fixed Answers
Based on Kant's dichotomy of judgment, I define the insurmountable decision-making boundary between humans and machines:
Determinant Judgment: Classifying special cases according to universal preset rules and outputting exclusive conclusions based on standardized paradigms. Its core is rule replication, standard implementation and case classification. AI achieves absolute superiority in this dimension with unprecedented accuracy and efficiency.
Reflective Judgment: The exclusive ultimate mental capacity of humans. Facing unprecedented scenarios without precedents, rules, data or standard answers, individuals must independently establish judgment criteria, define value weights, weigh right and wrong, and bear consequential responsibilities.
Rooted entirely in unique individual personality, conscience and life experience rather than collective paradigms and historical experience, reflective judgment represents the supreme manifestation of human subjectivity. AI can fit all knowns, yet it can never establish criteria for the unknown, generate value for blanks, or make autonomous choices in dilemmas.
Not an End, But a Beginning
AI will not end human civilization. It thoroughly exposes a core human civilization flaw: when instrumental cognition, labor, creation and execution are fully replaced by machines, humans must return to their irreplaceable essential values.
The automation of labor and work constitutes an ultimate coercion and invitation for humanity: after stripping all instrumental, repetitive and rule-based tasks, human core missions return to three exclusive capacities: independent thinking, value judgment, and public action.
All three capacities share the sole fundamental premise: an independent human individual capable of self-dialogue, self-reflection, interpersonal confrontation and active initiation.
AI remains eternally a instrumental "What", devoid of selfhood, conscience, speculation, judgment and initiative. Humans, as essential "Who", are innately endowed with the miracle of beginning, the freedom of speculation and the dignity of action — the eternal and insurmountable core foundation of human civilization.