AI as Archon: Algorithmic Governance and the Loss of Autonomy

Artificial intelligence is no longer only a tool that answers questions, sorts files, or generates images. Increasingly, it participates in systems that rank, predict, filter, recommend, police, approve, deny, optimise, and govern. In Gnostic language, this is why AI can be read as an archonic force: not because machines are demons, but because algorithmic systems often watch, weigh, classify, constrain, and administer human life without wisdom, transparency, or embodied responsibility.

In Plain Terms

To call AI an Archon is not to claim that artificial intelligence is literally a supernatural being. It is to use an ancient Gnostic pattern to understand a modern structure of power. In Gnostic texts, Archons are ruling powers: administrators of a lower order, forces that constrain perception, imitate higher truth, and maintain control through ignorance.

Modern AI systems can behave in a similar structural way. They classify people through data. They recommend what is seen. They hide what is not shown. They score risk. They generate synthetic content. They operate through opaque systems that even specialists may struggle to fully explain. They shape behaviour while appearing neutral.

The danger is not that AI is “evil”. The danger is that powerful systems can make decisions without understanding, accountability, compassion, memory, or conscience. The Gnostic warning is precise: administration without wisdom becomes a cage, even when the cage is efficient.

Primary Sources and Traditions Discussed

• Gnostic cosmology: Archons, rulers, false authority, ignorance, imitation, and the soul’s struggle for recognition.
• Nag Hammadi texts: especially the Apocryphon of John, Hypostasis of the Archons, and related writings on cosmic administration and spiritual awakening.
• Modern AI governance: algorithmic decision-making, recommendation systems, predictive analytics, synthetic media, AI agents, and automated public administration.
• Digital attention systems: recommender engines, infinite scroll, engagement loops, behavioural optimisation, and platform curation.
• AI risk and regulation: explainability, auditability, high-risk AI, human oversight, autonomous weapons, and emerging AI governance frameworks.
• Contemplative resistance: direct attention, analogue friction, interiority, discernment, and the refusal to reduce the human being to data.

How to Read This Article

This article uses Gnostic language symbolically and structurally. AI systems are not being presented as literal Archons from ancient myth. The claim is more subtle: the Archon is a pattern of rule without gnosis, and AI increasingly reproduces that pattern in technical form.

The point is not paranoia. The point is recognition. Algorithmic systems are built by humans, companies, states, institutions, and markets. They reflect incentives, training data, design choices, legal frameworks, and social priorities. When they are treated as neutral or inevitable, their power becomes harder to challenge.

Read “AI as Archon” as a diagnostic lens: a way to ask where technology assists human flourishing, where it silently governs, where it imitates understanding, and where it must be made accountable to human dignity.

Table of Contents

• Modern Companion: False Authority and Automated Judgement
• The Code That Rules
• The 2026 Landscape: Four Domains of Algorithmic Governance
• The Seven Parallels: Mapping Ancient Archons to AI Systems
• What Is New: The Specificity of AI Archons
• Explainable AI and the Demand for Gnosis
• False Authority and Automated Judgement
• Resistance: Gnosis Against the Algorithm
• The Question of Gnosis
• Related Glossary Terms
• Read Next
• Frequently Asked Questions
• Further Reading
• References and Sources

The Code That Rules

The shift from tool to governor did not happen in one dramatic moment. It happened by delegation. First the machine calculated. Then it ranked. Then it recommended. Then it screened, scored, predicted, flagged, filtered, matched, denied, escalated, and optimised.

A spreadsheet can assist a human decision. A scoring system can shape a human decision. A black-box model can quietly become the decision. That is the threshold this article is concerned with: the point at which technical assistance becomes administrative power.

By 2026, AI is woven into business operations, public services, content platforms, finance, hiring, health administration, surveillance, and security. McKinsey’s 2025 global survey reported that most organisations now use AI in at least one business function, while agentic AI is moving from experiment into organisational planning. The age of casual automation has become the age of delegated judgement.

The Gnostic lens becomes useful here because it asks a question that engineering alone does not answer: what happens when a system has power without wisdom? What happens when it administers without compassion, predicts without context, and imitates intelligence without self-knowledge?

The ancient Archon is not merely a monster. It is a ruler of a partial world that mistakes its partiality for total truth. That is the danger of algorithmic governance. The model sees what its data permits it to see. It ranks according to the objective it has been given. It optimises for the measurable. It may be useful. It may also be blind.

The algorithm does not need hatred to constrain. It only needs an objective function narrow enough to mistake the human being for a score.

The 2026 Landscape: Four Domains of Algorithmic Governance

Algorithmic governance does not appear in only one place. It spreads through domains where prediction, ranking, and optimisation promise efficiency. Four areas show the archonic pattern with particular clarity.

1. Predictive Public Administration: The Algorithmic Leviathan

Public bodies increasingly use data systems for fraud detection, resource allocation, eligibility checks, risk assessment, service delivery, and prioritisation. Some systems may reduce delays or expose patterns human administrators miss. Others can make harmful decisions harder to challenge because the logic is buried in software, vendors, data pipelines, or statistical assumptions.

Traditional bureaucracy can be slow, frustrating, and unjust. Algorithmic bureaucracy can be faster, less visible, and harder to appeal. The citizen is no longer only standing before an official. They may be standing before a model, a score, a threshold, or a rule they never see.

The archonic pattern is administration without encounter. A person becomes a risk profile. Need becomes a probability. Error becomes “system output”. The human face is replaced by a dashboard.

2. Algorithmic Policing and the Pre-Crime Paradigm

Predictive policing systems analyse historical crime data to forecast future risk. They may direct police resources toward locations, behaviours, or people marked as statistically significant. The problem is obvious once seen: historical data reflects historical policing. If a community has been over-policed, the data may treat that over-policing as evidence of future danger.

The model then sends more attention to the same places, generating more records, which reinforces the next prediction. The loop looks objective because it is numeric. But a number can still carry the scent of old injustice.

In Gnostic terms, this is the ruler mistaking the record of the lower world for truth itself. The system sees pattern, but not history. It sees correlation, but not suffering. It sees risk, but not the human story that produced the data.

3. Platform Governance: The Curated Self

Social platforms, search engines, short-form video feeds, and recommender systems govern attention at planetary scale. They do not usually control by direct command. They control by curation: this appears, that disappears; this becomes popular, that remains buried; this is frictionless, that requires effort.

The result is not only a curated feed. It is a curated self. The system predicts what you will click, then feeds you more of the predicted self, then trains the future prediction on your response. Identity becomes a loop between appetite and metric. The old Archons ruled the spheres. The new ones rule the scroll.

This is soft archonic power. It rarely says “you may not think this”. It simply arranges the world so that some thoughts never arrive with enough force to matter.

4. Autonomous Weapons: Lethal Delegation

The most extreme form of algorithmic delegation appears in military systems that identify, track, recommend, or potentially engage targets with varying degrees of human involvement. Autonomous weapons raise one of the starkest questions of the age: can the decision to kill be delegated to a system that does not know what death is?

The United Nations has recognised the urgency of the issue. In December 2024, the UN General Assembly adopted Resolution A/RES/79/62 on lethal autonomous weapons systems by 166 votes in favour, 3 against, and 15 abstentions. The resolution affirmed the relevance of international humanitarian law and called for continued international discussion.

The archonic danger here is not science fiction. It is responsibility dissolving into architecture. A commander blames the tool. A vendor blames the user. A state blames necessity. The system produces an outcome, but the moral subject disappears into the machinery.

The Seven Parallels: Mapping Ancient Archons to AI Systems

The parallels between ancient Gnostic Archons and modern AI systems should not be read as one-to-one identity. They are structural echoes. Each ancient function finds a technical analogue.

1. Watching: Total Data Collection

Ancient pattern: the rulers observe and administer the lower realm.

Modern analogue: surveillance capitalism, device tracking, location data, biometric systems, platform analytics, browser histories, smart homes, workplace monitoring, and behavioural data collection.

The digital watcher does not need a tower. It lives in phones, cameras, apps, cards, cookies, sensors, dashboards, and permissions clicked too quickly. The Panopticon has become pocket-sized and subscription-based.

2. Weighing: Risk Scoring

Ancient pattern: the soul is judged, measured, and sorted by powers that claim authority.

Modern analogue: credit scoring, risk assessment, fraud scoring, predictive policing, insurance pricing, hiring filters, welfare eligibility systems, and automated ranking.

The human being becomes a probability. The question “Who are you?” becomes “What does the model predict you will do?” That is a profound spiritual reduction, even when the mathematics is tidy.

3. Constraining: Algorithmic Curation

Ancient pattern: the rulers establish the boundaries of the possible and keep the soul inside a narrowed world.

Modern analogue: recommendation systems, ranking algorithms, search visibility, engagement optimisation, behavioural nudging, and filter bubbles.

The possible is not always banned. Often it is simply not recommended. What is never shown gradually becomes unthinkable. A prison does not need bars if the map never shows the road out.

4. Imitating: Synthetic Media

Ancient pattern: the archonic world imitates higher reality without possessing its living depth.

Modern analogue: deepfakes, synthetic voices, AI-generated images, AI-written text, artificial companions, simulated intimacy, and machine-generated authority.

Generative AI can be useful, creative, and beautiful. It can also produce the counterfeit at industrial scale: images without witness, speech without experience, intimacy without reciprocity, and knowledge-shaped language without knowing.

5. Ignorance: Black Box Opacity

Ancient pattern: the Archons do not know the fullness above them. They mistake their limited realm for total reality.

Modern analogue: opaque models, neural networks whose internal reasoning is difficult to interpret, proprietary systems, incomplete audit trails, and decisions that cannot be meaningfully explained to affected people.

The system can produce an answer without understanding the life it affects. It can classify without compassion, predict without wisdom, and optimise without knowing what should never be optimised.

6. Enforcement: Engagement Loops

Ancient pattern: the soul is returned to forgetfulness, habit, and lower identification.

Modern analogue: infinite scroll, notification loops, variable rewards, gamified metrics, streaks, outrage cycles, and attention traps.

The user is not forced to remain. The user is trained to return. This is subtler than command. It is captivity through appetite, polished smooth until the thumb keeps moving by itself.

7. Invisibility: Proprietary Systems

Ancient pattern: the rulers operate behind the curtain, known by effects more than by direct appearance.

Modern analogue: trade secrets, closed-source models, contractual opacity, private platform governance, invisible ranking criteria, and distributed accountability.

The system is everywhere and nowhere. No single person appears to govern, yet governance occurs. No one seems fully responsible, yet consequences land on real bodies. The Archon has discovered the corporate structure.

What Is New: The Specificity of AI Archons

AI governance repeats older patterns of bureaucracy, surveillance, and social control. But it also introduces new features that make the modern formation distinct.

Velocity and Scale

Human bureaucracy moves at human speed. It is often slow, inefficient, and frustrating, but its slowness creates friction. There is time to ask, appeal, resist, delay, misunderstand, reinterpret, or simply catch breath.

Algorithmic systems can operate at vast scale and speed. A model can sort thousands or millions of cases quickly. A platform can alter what entire populations see within moments. A recommender system can adjust its feed in real time. The gap between observation and response narrows until governance begins to feel atmospheric.

Adaptability and Learning

Older rule systems were comparatively static. AI systems can adapt from new data, changing behaviour as conditions shift. In some settings this is useful. In others, it means the constraint learns from resistance.

A user changes behaviour, and the model updates. A worker learns how to pass a filter, and the filter changes. A community resists visibility, and the system searches for new signals. The archonic structure becomes responsive.

Opacity and Inexplicability

Deep learning systems can be difficult to interpret. Some forms of AI produce outputs through complex internal relationships that do not translate easily into human-readable reasons. This creates a civic problem: people may be affected by decisions that cannot be explained in a way they can understand or challenge.

This is the modern form of Gnostic obscurity. A power operates. Its effects are real. Its reasoning is hidden. The affected person receives the verdict, not the path by which the verdict arrived.

Delegation and Diffusion

AI governance often diffuses responsibility across developers, vendors, deployers, datasets, public agencies, contractors, and users. When harm occurs, each layer can point elsewhere. The model did it. The vendor supplied it. The agency relied on it. The data trained it. The operator misunderstood it. The policy permitted it.

No one governs, yet governance happens. That is a signature of the distributed Archon: power without a face, decision without a visible decider, consequence without confession.

Explainable AI and the Demand for Gnosis

If opacity is one of the central problems, then explainability is one of the necessary responses. Explainable AI, algorithmic auditing, risk documentation, model cards, meaningful notices, contestability, and human oversight are not merely technical governance tools. In this symbolic frame, they are demands for gnosis: power must become knowable.

The European Union’s AI Act is important because it treats AI through a risk-based framework and places particular obligations on high-risk systems. Human oversight, transparency, record-keeping, and risk management are not decorative ideals. They are attempts to prevent automated systems from silently harming health, safety, and fundamental rights.

The NIST AI Risk Management Framework provides another major reference point, especially around identifying, assessing, managing, and governing AI risk. Such frameworks matter because they refuse the fantasy that technical power can regulate itself through good intentions alone.

Yet explainability has limits. A partial explanation can become a new mask. A compliance report can become a ritual of legitimacy. A human-in-the-loop can become a rubber stamp if the human lacks time, authority, training, or courage to override the system.

The Gnostic demand goes deeper: not only “explain the model”, but “show who benefits, who is harmed, who can appeal, who can refuse, who is accountable, and whether this system should exist at all”. Gnosis is not information. It is liberating recognition.

False Authority and Automated Judgement

AI becomes false authority when its outputs stop being treated as tools and begin to function as judgement. A risk score becomes a verdict. A ranking becomes reality. A recommendation becomes destiny. A chatbot becomes conscience. A dashboard becomes the face of the institution.

This is the bridge into Neo Gnosticism and False Authority. False authority is not only wrong information. It is the displacement of the faculty that would test information. In algorithmic life, that displacement can happen through speed, opacity, convenience and institutional habit: the machine says, the human rubber-stamps, and responsibility vanishes into the pipeline.

The issue is not whether AI can assist human judgement. It often can. The issue is whether the system is allowed to inherit authority it cannot carry. A model can classify, predict, rank and generate. It cannot possess conscience, suffer consequences, understand a life from within or practise gnosis. When its fluency is mistaken for wisdom, technical power begins to wear the robe of spiritual authority.

False authority begins when the score is treated as the soul.

ZenithEye editorial framing

The Neo Gnostic response is restored proportion. Use the model, audit the model, contest the model, but do not kneel to the model. The inner lamp by which authority is tested must remain human, embodied, accountable and awake.

Resistance: Gnosis Against the Algorithm

If AI governance repeats archonic patterns, resistance begins with recognition. Not panic. Not technophobia. Not romantic retreat from all tools. Recognition.

AI can be useful. Search can help. Translation can open texts. Accessibility tools can widen participation. Medical systems can assist clinicians. Pattern detection can reveal genuine risk. The issue is not whether technology should exist. The issue is whether technology remains a tool or becomes an invisible ruler.

Epistemic Resistance: Demand Transparency

Ask what system is being used. Ask what data trained it. Ask whether the output can be appealed. Ask who is accountable. Ask whether a human can override it. Ask whether affected communities were consulted. Ask whether the model’s purpose is aligned with human flourishing or only efficiency.

This is not technical fussiness. It is spiritual hygiene in a datafied world. The first move of the Archon is to appear inevitable. The first move of gnosis is to ask who built the throne.

To understand that you are being shaped by hidden systems is the first step toward refusing their claim to be reality itself.

ZenithEye editorial framing

Practical Resistance: Cultivate Friction

Algorithmic systems thrive on convenience. Friction is therefore a form of sovereignty. Use cash sometimes. Read physical books. Keep offline notes. Walk without tracking. Disable unnecessary notifications. Search outside your usual platform. Visit websites directly. Build bookmarks. Print what matters. Learn enough technical literacy not to be hypnotised by interfaces designed to hide their own machinery.

Friction does not mean rejecting modern life. It means refusing total capture. The system cannot optimise what it cannot measure. It cannot profile what you do not endlessly feed. It cannot govern the silence you keep intact.

Spiritual Resistance: Preserve Interiority

Contemplative practice protects what the feed cannot own. Sit without posting. Pray without metrics. Walk without recording. Read without extracting content. Let an experience belong to the soul before it becomes shareable material.

Direct experience is not anti-technology. It is pre-technology. It is the root layer from which wise use of technology becomes possible. Without it, the self becomes a reflection in the platform’s mirror, edited by engagement and fed back as identity.

The Question of Gnosis

We have built systems that watch, weigh, classify, predict, recommend, imitate, and decide. Some of these systems are useful. Some are dangerous. Most are mixed, because they are built inside mixed human societies by mixed human motives.

The Gnostic question is not simply whether AI is good or bad. The question is whether we remain awake inside the systems we build. Do we know what is shaping us? Do we understand the rulers we have installed? Can we distinguish assistance from governance, imitation from wisdom, efficiency from truth?

The Archon is defeated not by panic, but by recognition. Once a system is seen as a system, it loses the aura of inevitability. Once the ruler is recognised as partial, it can no longer pretend to be the whole. Once the human remembers the spark, the machine returns to its proper place: tool, not god; servant, not sovereign; mirror, not master.

---

---

Related Glossary Terms

These terms help clarify the Gnostic, technological, and political framework behind AI as Archon:

• Archons: ruling powers in Gnostic cosmology, often associated with false authority, limitation, ignorance, and cosmic administration.
• Gnosis: direct liberating knowledge or recognition, not merely information or belief.
• Demiurge: the lower creator or craftsman figure in some Gnostic systems, often linked with ignorance, imitation, and false sovereignty.
• Counterfeit Spirit: false imitation of authentic spiritual life, used in Gnostic texts to describe a force that mimics but does not liberate.
• False Authority: any teacher, system, institution, platform, algorithm or machine that replaces direct knowing with dependence, fear, obedience or borrowed certainty.
• Authority Capture: the gradual surrender of judgement, conscience, attention or direction to a system that benefits from dependence.
• Algorithmic Authority: authority granted to feeds, rankings, scores and automated systems when their arrangement of reality is mistaken for reality itself.
• Machine Authority: the tendency to treat automated systems as arbiters of guidance, truth or meaning because their outputs sound fluent, neutral or final.
• Spiritual Outsourcing: surrendering discernment to a tool, teacher, platform or system that cannot do the inner work of recognition for the seeker.
• Algorithmic governance: use of automated systems to classify, rank, predict, recommend, allocate, deny, or intervene in human life.
• Black-box AI: AI system whose internal reasoning or decision pathway is difficult for humans to understand or explain.
• Explainable AI: methods and governance practices intended to make AI outputs more understandable, auditable, and accountable.
• Predictive policing: use of data and models to forecast crime risk, often criticised for reinforcing historical policing bias.
• Surveillance capitalism: economic model that extracts behavioural data for prediction, influence, and profit.
• Recommendation system: algorithmic system that selects and ranks content, products, posts, videos, or search results for users.
• Engagement loop: design pattern that encourages repeated user return through reward, novelty, notification, and behavioural conditioning.
• Algorithmic sovereignty: the practice of retaining conscious agency, attention, and discernment in environments shaped by algorithms.

Read Next

For the strongest next step, continue into the companion article on false authority, machine judgement and the theft of direct knowing:

Neo Gnosticism and False Authority: Gurus, Algorithms and the Theft of Direct Knowing

This companion article follows the moment an AI system, platform, guru, metric or institution stops being a tool and starts being treated as the judge of what is real.

Then continue with Digital Suppression: Algorithmic Deplatforming and Modern Censorship, which examines how forbidden knowledge and institutional suppression have shifted into search visibility, platform ranking, shadow removal, demonetisation and attention control.

Frequently Asked Questions

Study Note: This article explores AI, algorithmic governance, surveillance, autonomous weapons, platform curation, false authority and Gnostic symbolism for educational and reflective purposes. It does not provide legal, technical, cybersecurity, employment, investment, or mental-health advice. If you are dealing with digital stalking, workplace monitoring, algorithmic discrimination, state surveillance, or legal harm from automated systems, seek qualified legal, technical, or professional support. For the wider spiritual-discernment layer, read Neo Gnosticism and False Authority. Digital minimalism, analogue resistance and privacy practices should be adapted to your actual circumstances, especially where employment, healthcare, safety or essential services are involved.

Further Reading

The following live ZenithEye links continue the themes of algorithmic control, digital suppression, archons, simulation, and modern Gnostic resistance:

• Neo Gnosticism and False Authority – Gurus, algorithms, AI advisers, authority capture and the recovery of direct knowing.
• Digital Archons: How Algorithms Shape Attention – The attention-level companion on feeds, rankings, recommendation systems and algorithmic authority.
• Digital Suppression: Algorithmic Deplatforming and Modern Censorship – How suppression has moved from visible censorship into search ranking, platform visibility, and algorithmic disappearance.
• Archons: The Ruling Powers That Shape Reality – The core Gnostic concept behind systems that administer, constrain, imitate, and obscure.
• Genie 3 and the Simulation Threshold: When AI Starts Building Worlds – AI world-generation and the shrinking distance between imagination, simulation, and rendered environments.
• The Living Thread: How Forbidden Knowing Survives the Fire – Historical transmission of forbidden knowledge through suppression, survival, and rediscovery.
• Gnosis in the Digital Age: Reclaiming Your Sovereignty from the Algorithm – Practical digital sovereignty, attention discipline, and spiritual autonomy amid algorithmic environments.
• Modern Resonances: Ancient Gnostic Wisdom for Today – How ancient diagnostic tools apply to twenty-first-century spiritual and technological conditions.
• AI as the New Demiurge: Yaldabaoth in the Digital Age – The cosmological implications of artificial intelligence, simulation, and counterfeit creation.
• Quantum Utility and the Glitch in the Matrix: Escaping Archonic Computation – Quantum indeterminacy, computation, and symbolic escape from deterministic control systems.
• Nag Hammadi Library: Complete Guide to the Gnostic Scriptures – The source archive for ancient Archon mythology and the wider Gnostic imagination.

References and Sources

The following sources support the Gnostic, technological, legal, and public-policy framework used in this article.

Gnostic and Historical Sources

• Apocryphon of John. Nag Hammadi Codex II,1; III,1; IV,1; Berlin Codex 8502,2.
• Hypostasis of the Archons. Nag Hammadi Codex II,4.
• On the Origin of the World. Nag Hammadi Codex II,5; XIII,2.
• Robinson, James M., ed. (1990). The Nag Hammadi Library in English. Revised edition. San Francisco: HarperOne.
• Meyer, Marvin, ed. (2007). The Nag Hammadi Scriptures. New York: HarperOne.
• Jonas, Hans. (1958). The Gnostic Religion. Boston: Beacon Press.
• King, Karen L. (2003). What Is Gnosticism? Cambridge: Harvard University Press.
• Williams, Michael Allen. (1996). Rethinking “Gnosticism”: An Argument for Dismantling a Dubious Category. Princeton: Princeton University Press.

AI Governance, Law, and Risk Frameworks

• European Union. (2024). Artificial Intelligence Act. Regulation laying down harmonised rules on artificial intelligence, including risk-based obligations for AI systems.
• European Commission. AI Act: Regulatory Framework for Artificial Intelligence. Official policy overview of the EU AI Act.
• National Institute of Standards and Technology. (2023). Artificial Intelligence Risk Management Framework (AI RMF 1.0). U.S. Department of Commerce.
• National Institute of Standards and Technology. Trustworthy and Responsible AI Resource Center. Implementation materials for the AI RMF.
• United Nations General Assembly. (2024). Resolution A/RES/79/62: Lethal Autonomous Weapons Systems. Adopted 2 December 2024.
• United Nations Digital Library. (2024). Lethal Autonomous Weapons Systems: Resolution Adopted by the General Assembly. Official record for A/RES/79/62.

AI Adoption, Black-Box Systems, and Explainability

• McKinsey and Company. (2025). The State of AI: Global Survey 2025. QuantumBlack AI by McKinsey.
• IBM. (2024). What Is Black Box AI and How Does It Work? IBM Think.
• Rudin, Cynthia. (2019). “Stop Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead.” Nature Machine Intelligence, 1, 206-215.
• Burrell, Jenna. (2016). “How the Machine Thinks: Understanding Opacity in Machine Learning Algorithms.” Big Data & Society, 3(1).
• Doshi-Velez, Finale, and Kim, Been. (2017). “Towards A Rigorous Science of Interpretable Machine Learning.” arXiv.

Surveillance, Platforms, and Attention

• Zuboff, Shoshana. (2019). The Age of Surveillance Capitalism. New York: PublicAffairs.
• Pariser, Eli. (2011). The Filter Bubble: What the Internet Is Hiding from You. New York: Penguin Press.
• Noble, Safiya Umoja. (2018). Algorithms of Oppression: How Search Engines Reinforce Racism. New York: NYU Press.
• Eubanks, Virginia. (2018). Automating Inequality: How High-Tech Tools Profile, Police, and Punish the Poor. New York: St. Martin’s Press.
• O’Neil, Cathy. (2016). Weapons of Math Destruction. New York: Crown.
• Pasquale, Frank. (2015). The Black Box Society: The Secret Algorithms That Control Money and Information. Cambridge: Harvard University Press.
• Citron, Danielle Keats. (2008). “Technological Due Process.” Washington University Law Review, 85(6), 1249-1313.

Autonomous Weapons and Military AI

• International Committee of the Red Cross. Autonomous Weapon Systems and International Humanitarian Law. Policy and legal materials on autonomy in weapons systems.
• Human Rights Watch. (2024). “Killer Robots: UN Vote Should Spur Treaty Negotiations.” Statement on UN General Assembly Resolution 79/62.
• U.S. Department of Defense. (2023). Directive 3000.09: Autonomy in Weapon Systems.
• Campaign to Stop Killer Robots. Public materials on lethal autonomous weapons and international regulation.