By Efstathios Iliopoulos and Beatrice Ducci
In the regime of ancient Athens and later in Rome, slavery formed an important component of economic and social organization, although these societies were not sustained exclusively through enslaved labor. In Athens, productive activity also depended heavily on small farmers and metics[1]. Nevertheless, enslaved populations occupied a central role in domestic service, agriculture, mining, and craft production, allowing many citizens to dedicate greater time to political participation, military obligations, philosophy, and cultural life[2]. The historical significance of slavery therefore lies not only in its economic function, but also in what it reveals about a recurring human tendency: the desire to transfer what is considered dull, repetitive, or dangerous labor onto others in pursuit of freedom, status, or intellectual development.
This tendency did not emerge solely from cruelty or economic opportunism. Ancient societies operated under material limitations that made survival labor-intensive and uncertain. Agricultural production required immense physical effort, mining conditions were often life-endangering, and transportation systems depended almost entirely on human and animal exertion. Within such conditions, slavery became normalized as one among several mechanisms through which societies redistributed labor burdens. Perhaps the earliest documented instance of this structural redistribution are the mines of Laurion in Attica, where enslaved laborers worked under lethal conditions, while the Athenian citizens, who profited from them, never even visited them[3].
Yet normalization did not imply moral unanimity. Even in antiquity, the institution generated tensions and criticisms, especially among certain Stoic thinkers who emphasized the moral equality of human beings despite legal status differences[4]. The existence of slavery was therefore historically widespread, but never philosophically uncontested. The most explicit and emotionally direct indictment of the master-slave psychology belongs to Seneca (c. 4 BC–65 AD). In Epistulae Morales, Letter 47 – De servis (“On Slaves”) – he writes to his friend Lucilius urging him to treat enslaved persons as human beings, reminding him that fortune alone separates the free from the enslaved: “Ille servus est – sed fortasse liber animo” (“He is a slave – but perhaps he is free in spirit”)[5]. Seneca insists, with striking modernity, that masters who dehumanize their slaves ultimately dehumanize themselves. This observation mirrors the central concern of the present article: that systems of asymmetric service corrupt the moral psychology of those who rely upon them, whether the “master” is an Athenian citizen, a Roman senator, or a twenty-first century technology company.
The Roman world expanded slavery into a far larger and more economically integrated institution. Following Rome’s military expansion during the late Republic, vast agricultural estates (latifundia) increasingly depended on enslaved populations captured through conquest campaigns, transforming slavery into a deeply integrated component of Roman economic organization. Roman slavery permeated agriculture, mining, administration, domestic life, education, and large-scale construction projects. Yet Roman slavery was not socially uniform. Conditions varied dramatically depending on occupation, geography, and ownership. Agricultural and mining slaves frequently endured extreme brutality and short life expectancies, whereas some educated Greek slaves worked as tutors, accountants, physicians, or secretaries and could occasionally achieve manumission and limited social mobility[6]. This diversity complicates simplistic portrayals of Roman slavery while simultaneously illustrating how systems of domination adapt themselves across different economic functions.
Despite these variations, slave systems in both Greece and Rome shared a deeper structural logic: they externalized undesirable labor so that privileged groups could pursue activities considered more valuable or civilized. Leisure, political participation, military strategy, education, and artistic production became associated with social prestige precisely because others absorbed the burdens of necessary labor. Historians of political thought have noted that classical ideals of citizenship and intellectual cultivation were inseparable from these underlying economic arrangements[7]. In this sense, slavery was not merely an isolated institution, but a part of a broader civilizational framework that offered freedom for some to the detriment of others.
The structural relationship between servitude and civilization has also been acknowledged by contemporary political philosophers. Giorgio Agamben notably observes that “slavery is to ancient humanity what technology is to modern humanity: both, as bare life, watch over the threshold that allows access to the truly human condition”[8]. This comparison does not imply a moral equivalence between enslaved persons and technological systems. Rather, it highlights a recurring historical pattern in which societies externalize necessity, labor, and material dependence in order to sustain forms of political, intellectual, and cultural life considered distinctively “human”. Contemporary developments in artificial intelligence and robotics may therefore represent a technologically transformed continuation of this civilizational tendency, rather than a historical rupture.
Modern developments in robotics and artificial intelligence reveal a “technologically transformed” version of this same aspiration. Contemporary societies increasingly seek to automate forms of labor considered monotonous, dangerous, exhausting, or psychologically undesirable. Warehouse logistics, industrial manufacturing, customer support, transportation, data processing is progressively delegated to intelligent systems. However, the comparison with ancient slavery must be approached carefully: current AI systems are not conscious beings, do not possess subjective experience, and cannot suffer in the human sense. Nevertheless, one could argue that the analogy remains philosophically useful because it does not concern equivalence between humans and machines, but rather recurring patterns in how societies attempt to externalize labor and maximize convenience. What connects these historical and modern systems is less the status of the laborer than the psychology, the scope and the ethics of the master. In antiquity, this transfer occurred through enslaved populations. In industrial modernity, it occurred partly through mechanization and colonial extraction. In the emerging AI era, it increasingly occurs through algorithmic and robotic systems designed to function with uninterrupted obedience. The danger lies not in machines themselves, but in the social habits cultivated through relationships based entirely on utility, control, and dependence.
Aristotle himself anticipated the relationship between technological substitution and slavery in Politics Book I. Discussing the role of instruments and labor, he famously observed that “if every instrument could accomplish its own work” masters would no longer require slaves (Politics I.4, 1253b)[9]. Although formulated within the technological limitations of antiquity, Aristotle’s reflection reveals an early awareness that systems of servitude were historically connected to the absence of autonomous productive technologies. Contemporary developments in robotics and artificial intelligence therefore raise a question that Aristotle himself indirectly anticipated: if intelligent systems increasingly perform labor once assigned to humans, what social and political transformations might follow from this shift?
Historical slave societies often reshaped moral perception around hierarchical thinking. When groups become accustomed to viewing others primarily as instruments of productivity, empathy can narrow and domination can become normalized beyond its original context. Roman society, for example, developed rigid social stratifications that extended not only between slave and free, but also among citizens themselves through class and patronage systems. Seneca’s warning in Letter 47 that the practice of treating others as mere instruments degrades the practitioner – acquired a particular urgency in a society where such habits had become structural. His insight anticipates what contemporary scholars of AI governance describe as the risk of “moral distancing”: the progressive erosion of individual accountability as decisions are delegated to opaque computational systems[10]. For example, recent controversies involving algorithmic bias in predictive policing, automated hiring systems, and AI-assisted welfare allocation have demonstrated how institutions may deflect responsibility by attributing discriminatory or harmful outcomes to supposedly neutral technological processes. Similarly, the use of Lethal Autonomous Weapons Systems (LAWS) in military targeting raises unresolved questions regarding who remains legally and morally accountable when algorithmic decisions result in civilian harm.
Furthermore, contemporary AI-driven economies risk intensifying inequalities between those who own technological infrastructures and those displaced by them. Automation may liberate certain individuals from labor while simultaneously concentrating wealth, data, and decision-making power into increasingly centralized institutions[11]. Another major concern involves moral distancing and responsibility. Ancient slave owners could psychologically detach themselves from the suffering sustaining their lifestyles because exploitation had become socially “accepted”. Modern algorithmic systems may produce a comparable form of ethical distancing. Decisions involving employment, policing, healthcare, warfare, finance, and social surveillance are increasingly delegated to opaque computational systems. Scholars of AI governance warn that human oversight may become symbolic rather than substantive if decision-making processes exceed human understanding or accountability structures remain weak[12]. In such cases, institutions may evade moral responsibility by attributing controversial outcomes to technological systems rather than human choices.
There also exists a deeper philosophical contradiction in humanity’s pursuit of total liberation from labor. While labor has historically involved hardship and exploitation, it has also contributed to meaning, discipline, solidarity, technical mastery, and social participation. It also made art, especially literature, shine. Ancient elites often justified slavery as a pathway toward higher civilization and intellectual flourishing, yet Greek and Roman societies remained marked by inequality, revolt, and ethical contradiction. Likewise, a future centered entirely around automated convenience may risk weakening forms of participation that help sustain human purpose and collective responsibility. A society organized solely around comfort and optimization may gradually erode the civic and interpersonal bonds necessary for democratic life.
Ultimately, the analogy between ancient slavery and modern AI should not be understood as a literal equivalence between enslaved humans and artificial systems. Rather, it highlights a recurring civilizational pattern: humanity continually attempts to displace vulnerability, repetition, and necessity onto an external “other” in order to preserve freedom and aspiration for itself. The historical lesson of slavery is therefore not only about oppression, but also about the moral dangers that emerge when societies become excessively dependent on systems of asymmetrical service. The greatest risk posed by advanced AI may not be that machines become too human, but that human societies may increasingly construct new forms of dependency, passivity, and concentrated power around automated infrastructures.
[1] Resident foreigners who contributed substantially to trade and skilled labor.
[2] Cartledge, P. (2011). Ancient Greece: A history in eleven cities. Oxford University Press.
[3] William T. Loomis, (1998). Wages, Welfare Costs and Inflation in Classical Athens. University of Michigan Press.
[4] Garnsey, P. (1996). Ideas of slavery from Aristotle to Augustine. Cambridge University Press.
[5] Seneca, Letters on Ethics (Epistulae Morales). Letter 47. Trans. M. Graver & A.A. Long. University of Chicago Press, 2015.
[6] Scheidel, W. (2010). Slavery in the Roman economy. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.1663556
[7] Ambler, W. (1987). Aristotle on nature and politics: The case of slavery. Political Theory, 15(3), 390–410. https://doi.org/10.1177/0090591787015003007
[8] Agamben, G. (2015). The use of bodies, A. Kotsko (trans). Stanford University Press.
[9] Aristotle. Politics. Book I, 1253b32–1254a1. Translated by C.D.C. Reeve. Indianapolis: Hackett Publishing, 1998.
[10] Seneca, Ep. 47.11 Floridi, L., & Cowls, J. (2019). A unified framework of five principles for AI in society. Harvard Data Science Review, 1. https://doi.org/10.1162/99608f92.8cd550d1
[11] Brynjolfsson, E., & McAfee, A. (2014). The second machine age: Work, progress, and prosperity in a time of brilliant technologies. W. W. Norton & Company.
[12] Floridi, L., & Cowls, J. (2019). A unified framework of five principles for AI in society. Harvard Data Science Review, 1. https://doi.org/10.1162/99608f92.8cd550d1
