The Polyopticon: Data Gathering and State Technopower

A panopticon-inspired prison. Photo Credit: Wikimedia Commons.


The world is on the cusp of a fourth industrial revolution.[i] This revolution is driven by the increased connectivity, reactivity, and converging nature of modern technologies. It is enabled by constant data-streams analyzed through ultramodern techniques, which allow vendors to immediately integrate analysis into product and service development with minimal delays and to maximum effect. This process results in iterative development and design, updating every successive product or service to better meet evolving market demands. The relationships among producers, consumers, the market, and ‘final’ products will fundamentally change, creating mutually-influencing networks that will benefit integrated commercial enterprise to a greater degree than ever before. 5G’s massively-broadened bandwidth, reliability in data collection, and resolution of real-time data will enable faster, more reactive innovation and design.

Governments, too, might adapt and synergize the coexistence of decentralized connected data streams and analytical technologies to increase their own reactivity. The fourth industrial revolution is a model for the future of massive and individualized surveillance, analysis, and reactivity. The market has already normalized transponders, cameras, and microphones in billions of pockets and on billions of desks across the world.[ii] The major problem facing authorities interested in real-time surveillance is no longer the propagation of hardware but the centralization and analysis of surveillance data.[iii] Big data requires big analysis, and this is where the AI-revolution can support authorities’ efforts in ways that human-only analysis never could. This process—the marriage of massively-disseminated surveillance hardware streaming through modernizing super-high bandwidth infrastructure to AI-assisted analytics for the purpose of creating actionable knowledge, and all the policies required to make that happen—is the Polyopticon.

How Statistics Leverage State Power and the Tension Between Discipline and Citizenship

The Polyopticon is a play on the panopticon—a cylindrical prison in which a beehive of cells opens inward to a central guard tower, creating an architecture in which no single prisoner is ever assured of their privacy regardless of how lax surveillance over them might be.[iv] For many reasons, the panopticon has an incredibly limited application for governance over large populations, not least because it requires the literal reconstruction of physical reality. The panopticon is expensive, unscalable, and therefore impractical for governance. It underscores the classic constraints to states’ intelligence activities, namely: hardware, human capital, and scalability. But the Polyopticon, as a technopolitical tool, leverages various existing technologies to provide states with surveillance data streams and the tools required to convert them into actionable statistics that could not have existed even a decade ago.

Innovation in technology will not change the core elements of sovereignty: exercising a monopoly of violence among a population within a territory.[v] Much as hardware innovations alone have never been enough to create more actionable intelligence, innovations in violence have never been sufficient to secure sovereignty. States are always pressured to use the violence—and all the resources of their population and territory—available to them strategically, requiring them to match violence with knowledge of their problems and possible solutions. In their quest to build better strategies and leverage the resources available to them, states invented one crucial science: statistics.[vi] As Michel Foucault argued, the majority of statistics were aimed at leveraging biopower—the biologically-bound resources of the state such as human bodies, crops, and the energy extractable from various fuels available within a state—to expand the domestic and geopolitical reach of ‘state forces’ over more people, territory, and time.[vii] Whether they create positive externalities or not, biopolitics (policies and politics introduced to leverage biopower) like the census, public health programs, and personalized identification always serve to create or multiply the capabilities and statistics states see as key to increasing the projection of their power domestically and abroad for both high- and low-political aims.[viii]

Statistics, as Foucault explains, were married with the scientific process to create institutions and systems that simultaneously amplified the capabilities of the state and controlled human populations by subjecting them to discipline.[ix] Discipline was disassociated from biopolitics as populations wrested control—through what Foucault terms ‘counter-conduct’—from states by demanding rights, expressing freedoms, and normalizing citizenship.[x] State-discipline no longer forced resource productivity (human labor) as citizenship and free enterprise incentivized production and the creation of wealth, which themselves multiplied state-power. Discipline, instead, was directed at 1) leveraging internal resources in times of crisis (i.e., the draft, rationing, and nationalization) and 2) deviant behavior (i.e., criminality, revolutionary politics, non-conformity, treason).[xi] This tension, between the disciplining tendencies of the state and the liberalizing demands of citizenship exists to this day. The Polyopticon requires coordination—the art of state-strategy—and abets state-directed discipline by providing it with the actionable statistics it needs to surveil, analyze behavior, and identify deviant individuals, movements, and populations.

The Polyopticon

The individual pieces of the Polyopticon were not conspiratorially created, disseminated, or activated. The devices that can stream data in real-time, the ultramodern networks that connect them, the analytical techniques that synthesize knowledge from data were not devised for the purpose of surveilling, analyzing, or controlling populations and individuals. Smartphones, fiberoptic cables, web servers, big-data crunching, and artificial intelligence were all created to solve various commercial, private, and public problems. The fourth industrial revolution depends on the synergy of these technologies to create better services and products, faster. The Polyopticon will likewise leverage the same technologies to more rapidly create better actionable statistics.

This sounds innocuous enough, but to put it sarcastically: What have states ever done with more knowledge? Let it lie and gather dust? No. States utilize knowledge. As Foucault points out, knowledge—especially knowledge about a subject that the subject itself does not know—is an especially potent weapon against enemies both foreign and domestic.[xii]

A description—detailed though not technical—of how the Polyopticon turns surveilled data into actionable statistics clarifies the above argument. It is especially useful to add the additional frame of how the Polyopticon solves the three core problems of state-intelligence gathering: hardware, human capital, and scalability.

First, in search of actionable knowledge, the state must surveil. Here the Polyopticon taps into personalized devices and technologies, voluntarily employed across a broad spectrum of private, corporate, and public sectors. Smartphones with microphones, cameras, and transponders; electronic communications; web and cloud-based information storage; social and commercial media—all of these are sources of surveilable data once connected. States leverage mundane technologies that are nonetheless integral to modern life. Liberal and authoritarian governments alike might access these data-sources. In this way, the Polyopticon records citizens’ behavior within reality rather than reconstructing reality to influence behavior.

More conspicuous, of course, are countries in which governments purposefully place data-collecting technology around the lives of their citizen-subjects. The cheapness and connectivity of surveillance cameras has made them a ubiquitous part of urban life, and a realistic surveillance tool in Xinjiang.[xiii] Additionally, there are endless subroutines, the invisible-to-the-laity protocols that commercial interests or states can imbed into the digital and cyber lives of their customers/citizen-subjects to track their correspondence, movement, secrets, and personal lives like never before.[xiv] Again, billions voluntarily participate in the cybersphere, and many modern aspects of life—from banking to communication—practically require some degree of digital participation. Where biopolitics leveraged the nature of human biology to subject human bodies (to surveillance, statistics, discipline), technopolitics like the Polyopticon instead leverage ubiquitous technologies to the same purpose. Today’s surveillance hardware is cheap, disseminated, and so ubiquitous that it creates its own problem: volume.

Any armchair strategist knows an up-to-date picture of a situation can make the difference between the right decision and the wrong ones—and pictures today are more accessible from armchairs than ever before. 5G specifically—as infrastructure for highspeed communication and access to repositories and real-time sources of information—is a game-changer. It can handle the volume of real-time data communication and storage that the Polyopticon’s surveillance hardware gathers. The relay time from peripheral event of consequence to central state responder is shorter than ever before. Moreover, real-time data means final statistics that are more reflective of the actual, rather than the estimated, situations states want to influence. Utilizing high-volume networks to gather and centralize surveillance-data for further analysis is not even the most conspiratorial Polyopticon approach. China’s Huawei, for example, is a government-backed technological giant that can offer the ‘full-suite’ of 5G infrastructure and servicing, already synergizing and expediting the adoption of next-generation telecommunications that will enable the Polyopticon and further empower the state.[xv] Moreover, the exportation of state-backed 5G technologies to foreign states puts the surveil-able data within their territories at even greater risk of extraction to a different central authority who might use that data for nefarious purposes.[xvi] That data alone, even unanalyzed, could still be a goldmine for those states already preparing national Polyopticon strategies.

Centralized data is not knowledge. The utility of any intelligence apparatus is its ability to glean intelligence from surveillance. Here, there is as real a human capital gap between what states have and what they need, as there is a bandwidth gap between the first cellular networks and 5G. There is so much data gathered today—‘big data’ already alluded to—that unassisted humans could never make use of all of it, rendering the previous parts of the Polyopticon impractical, if not useless. This is where the most revolutionary of the Polyopticon’s synergized technologies comes in: machine learning and artificial intelligence. AI, as far as the Polyopticon is concerned, is useful because it can analyze patterns—and therefore identify deviances. The key hurdle is turning surveillance into data that AI can parse, and teaching AI to parse that data quickly and into actionable intelligence. What does that entail?

First it means creating AI(s) that can learn from a variety sources: sound, image, location, text, digital and analog communication, etc. Then it means feeding AI that data and teaching it to parse accurately.[xvii] This is a tedious task requiring constant human intervention. Various machine learning techniques can speed up that process of trial and error; however, success and failure still require human-assisted audit. Some nations, however, have a lot of humans who can be set to the task of auditing AI, bringing closer the day when AI can actively be relied upon to analyze surveillance data. AI-data farms use globally-sourced data as a goldmine from which to extract the nuggets that will evolve their AI faster than others’.[xviii] This sparks an especially profound realization: no longer is any data, gathered off any device anywhere definitely innocuous. Someone storing their morning run from a smartwatch to the cloud might unknowingly provide data later applied to teach an AI which one day could be used by authoritarian governments to sift through the activities of their citizen-subjects in search of deviant behaviors. It is no longer possible to rationalize that any data is innocuous.

The Polyopticon and Counter-Conduct

As already explained, authoritarian governments are advantaged in employing the Polyopticon because they face less domestic scrutiny and fewer governance barriers generally. Liberal states, however, can also leverage Polyopticon policies without introducing state-sponsored hardware by exploiting disseminated dual-use technologies. Where authoritarians can source AI-teaching data globally, liberal states (and corporations in them) face privacy and legal hurdles here, too. But once any state deploys the Polyopticon, those technologies can be copied by others—and this conduct will require counter-conduct to combat states’ tendency to discipline. Vigilance is the age-old response to surveillance, and a key piece to safeguarding citizens’ rights in an age when states might know more about their collective citizenry than the citizens know about themselves. When the Polyopticon might be used to disseminate disinformation, discourse centered in civil society combats nefarious narratives. Finally, as liberal governments are pressured to resort to the Polyopticon for security reasons, it is the electorates’ responsibility to remind them that changing behaviors on a national level is the exact definition of one state exercising power over another.


[i] For an overview of the fourth industrial revolution see: Klaus Schwab, The Fourth Industrial Revolution (London: Portfolio Penguin, 2017).

[ii] John Gilliom and Torin Monahan, SuperVision: An Introduction to the Surveillance Society (Chicago: University of Chicago Press, 2013), 11-26.

[iii] Sue Halpern, “The Terrifying Potential of 5G Technology,” The New Yorker, April 30, 2019,

[iv] Michel Foucault, “Discipline & Punish – Panopticism,” Michel Foucault, Info.,

[v] Daniel Philpott, “Sovereignty,” Stanford Encyclopedia of Philosophy, March 25, 2016,

[vi] Michel Foucault, Security, Territory, Population: Lectures at the Collège De France, 1977-1978, ed. Michel Senellart (New York, NY: Palgrave Macmillan, 2007), 100-101.

[vii] Ibid., 315.

[viii] Ibid., 323-26.

[ix] Ibid., 340-41.

[x] Ibid., 355-57.

[xi] Ibid., 353-54.

[xii] Ibid., 275.

[xiii] Chris Buckley and Paul Mozur, “How China Uses High-Tech Surveillance to Subdue Minorities,” The New York Times, May 22, 2019,

[xiv] David Lyon, Surveillance as Social Sorting: Privacy, Risk, and Digital Discrimination (London: Routledge, 2008), 13-30.

[xv] Elsa B. Kania, “Why China’s Military Wants to Beat the US to a Next-Gen Cell Network,” Defense One, January 08, 2019,

[xvi] Justin Sherman and Robert Morgus, “Authoritarians Are Exporting Surveillance Tech, And With it Their Vision for the Internet,” Council on Foreign Relations, December 5, 2018,

[xvii] CGP Grey, “How Machines Learn,” YouTube, December 18, 2017,

[xviii] VICE News, “China’s Big AI Advantage: Humans,” YouTube, May 31, 2019,

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.