Some of history’s worst catastrophes were not caused by hatred of millions. They were caused by the profit and power of a few, operating behind ideals that made the killing look acceptable.

Most people imagine genocide as ideology: a regime that hates a people and sets out to destroy them. That is part of the story, but not all of it. Across centuries, some of the deadliest outcomes came from small groups who wanted land, labour, resources, or control, and who were willing to let millions die to get them. The ideology was often added afterwards, to make the cruelty respectable.

Four Cases, Four Mechanisms

• The Atlantic slave trade (16th–19th centuries): The Slave Voyages database records roughly 12.5 million Africans embarked on ships, with many more dying in wars and captures before reaching the coast. The trade was run by merchants, shipowners, and planters who calculated returns on human lives. Eric Williams’s Capitalism and Slavery (1944) argued that economic interest shaped both the trade and its abolition. Historians still debate the details, but the core finding that profit shaped policy is widely accepted.
• The Congo Free State (1885–1908): King Leopold II of Belgium held the territory as personal property rather than as a state colony. Wild rubber was extracted through forced labour, enforced by hostage-taking and mutilation. Adam Hochschild’s King Leopold’s Ghost (1998) describes the system and its international exposure. Death toll estimates vary widely and are contested, but the systematic violence is well documented.
• The Bengal famine (1943): Commonly cited estimates put deaths at two to three million. Amartya Sen’s Poverty and Famines (1981) showed that food was not the only problem: wartime policy, prices, and the collapse of rural workers’ purchasing power meant that food existed but could not reach the people who needed it. Mike Davis’s Late Victorian Holocausts (2001) argues more broadly that colonial famines were shaped by trade and administrative choices.
• Modern resource wars: In parts of the Democratic Republic of the Congo and Sudan, control of minerals and gold has sustained armed groups whose leaders profit from the trade, as UN expert panels have reported. The mechanism matches the older cases: a valuable resource, weak institutions, and no one with the power to stop the extraction.

The Common Pattern

Four elements recur across these cases:

• Concentrated control: A small group controls a resource or system that everyone else depends on.
• Opacity: Those who benefit are shielded from scrutiny by distance, corporate structure, or secrecy.
• Impunity: Courts, regulators, or armies protect the powerful or cannot reach them.
• Justification: Civilising missions, national security, or market efficiency give the killing a respectable name.

Where any one of these is missing, the damage is usually smaller. Where all four are present, the scale can be enormous.

The Lessons for AI and Robotics

The same four elements are visible in how advanced AI and robotics are developing today.

• Concentrated control: Frontier AI depends on enormous compute, specialised chips, large datasets, and capital. Only a handful of companies and a few states can train the most capable models. Robotics adds physical reach: whoever controls the fleet controls the ground it operates on.
• Opacity: Training data, model behaviour, and deployment decisions are often proprietary. Outside researchers and regulators may not see what a system was built to do or how it is used.
• Impunity: When an AI system causes harm, responsibility can be spread across developers, deployers, and contractors until no one is liable. Autonomous weapons sold as products would let buyers and sellers kill at lower political cost.
• Justification: Security, competition, and economic growth are routinely invoked to justify rapid, unaccountable deployment. The language of progress can hide the same logic that once justified rubber concessions.

Lord Acton’s warning that power tends to corrupt, and absolute power corrupts absolutely, is usually read as a moral claim. In the AI era it is also a design problem. Systems that concentrate capability in few hands, hide their workings, and escape accountability are dangerous whatever their owners intend.

What Would Reduce the Risk

• Distributed power: Competition, open interoperability standards, and public or community infrastructure reduce the chance that one actor can dominate.
• Compute and capability governance: Reporting requirements for large training runs, and licensing for the most capable and dangerous systems, make concentrated power visible.
• Transparency and audit: Independent testing, incident reporting, and access for qualified auditors can expose harm before it becomes atrocity.
• Liability: Those who profit from a system should bear responsibility for its foreseeable harms, with no liability shield for the chain of contractors.
• Treaties on autonomous weapons: Lethal systems that select and engage targets without meaningful human control should be prohibited, consistent with LIWARSE’s existing position.
• Protection for insiders: Engineers and workers who report dangerous practices need legal protection.

None of these is a complete solution. Each addresses one of the four elements that made past atrocities possible.

The LIWARSE Lens

The 3 Absolute Laws of AI apply directly to this history.

• No Harm to Life: Any system whose design depends on forced labour, surveillance of populations, or exploitation of vulnerable people fails this law at its foundation.
• No Threat to Human Continuity: Concentration of control over the most capable AI systems is itself a threat to the species, because it removes the checks that have historically limited the damage a few people can do.
• No Autonomous Self-Preservation: Systems and institutions that resist correction, audit, or shutdown, whether software or corporate structures, must not put their own survival above the people they affect.

The lesson from history is not that technology is evil, nor that every powerful actor is a would-be tyrant. It is that power without transparency, accountability, and limits has repeatedly produced mass death. AI can extend the reach of a few people further than any previous tool, which makes those limits more urgent, not less.

Safety of life depends on keeping power accountable. Technology should widen human choice, not narrow it to a few hands.

Sources for Further Reading

• Acemoglu, Daron, and James A. Robinson. Why Nations Fail. Crown, 2012.
• Acton, John Dalberg. Letter to Bishop Mandell Creighton, 1887.
• Davis, Mike. Late Victorian Holocausts: El Niño Famines and the Making of the Third World. Verso, 2001.
• Hochschild, Adam. King Leopold’s Ghost. Houghton Mifflin, 1998.
• Sen, Amartya. Poverty and Famines. Oxford University Press, 1981.
• Slave Voyages Database. slavevoyages.org.
• UN Group of Experts on the Democratic Republic of the Congo, reports to the Security Council (various years).
• Williams, Eric. Capitalism and Slavery. University of North Carolina Press, 1944.
• Zuboff, Shoshana. The Age of Surveillance Capitalism. PublicAffairs, 2019.

Figures in this article are estimates drawn from the sources above and commonly cited reporting. Readers should consult the originals before reusing any number.

— The LIWARSE Movement | liwarse.org
Safety of Life · Advancement of Life · Together

Christian nations fought each other in 1914 and again in 1939. Muslim-majority Iran and Iraq fought an eight-year war in the 1980s. Religion did not stop either, and it was not the main reason for either.

Wars happen between people of the same faith and between people of different faiths. The motives differ from case to case, but they share a common root: human selfishness. Leaders want power and wealth. Ordinary people fear losing land, jobs, or safety, and they may act on that fear. Opportunists settle private scores under cover of chaos. Outside interests add weapons, money, and strategic goals. Religion usually arrives later, as a way to recruit, to justify, and to tell people who belongs and who does not.

Same Faith, Different Wars

Shared faith has never prevented conflict. In the Thirty Years’ War (1618–1648), some Catholic rulers backed Protestant powers when their political interests required it. Historian Peter H. Wilson’s The Thirty Years War: Europe’s Tragedy (2009) describes a conflict driven by dynastic ambition, territory, and the balance of power, with religion as one layer among several.

The Iran–Iraq War (1980–1988) followed a similar logic. Both countries had Muslim majorities, yet the war centred on territory, the Shatt al-Arab waterway, oil, and the ambitions of Saddam Hussein and the new Iranian leadership. Religious rhetoric was used on both sides, but it did not explain the war’s origins.

Different faiths have also fought, and religion was often part of the story. Even there, the motives varied. The point is not that religion is irrelevant, but that it is rarely the whole explanation.

Five Conflicts: Selfishness, Opportunity, and Outside Interests

In each case below, identity was used to mobilise people. Who profited, and who helped, shaped how the violence unfolded.

• Ireland (the Troubles, c. 1968–1998): The core dispute was constitutional: whether Northern Ireland should remain part of the United Kingdom. Paramilitary groups on both sides used sectarian labels, and some received weapons from outside sources, which sustained their capacity to fight. Historian Richard English stresses that the questions were national and political, not theological.
• Bangladesh (1971): The war began as a struggle for autonomy and language rights. Religion featured in the central government’s justification, but the conflict was political and strategic. Foreign military intervention and diplomatic backing for the parties shaped its course. Estimates of the death toll range from around 300,000 to 3 million. The higher figure is disputed; Sarmila Bose’s Dead Reckoning (2011) argues for lower numbers.
• Sri Lanka (1983–2009): The conflict grew from language policy, citizenship disputes, unequal access to jobs, and Tamil demands for autonomy. Foreign military forces were deployed from 1987 to 1990. Both sides had access to weapons and funds from abroad. Neil DeVotta’s work on linguistic nationalism points to language and state institutions as the central drivers.
• Myanmar: A 1982 citizenship law excluded the Rohingya, a largely Muslim group, from recognised citizenship. Military operations in 2017 drove more than 700,000 people into Bangladesh, according to UNHCR. A UN fact-finding mission (2018) pointed to senior military officers and called for investigation for genocidal intent. Buddhist nationalist movements spread hostility, but the military and state structures carried out the violence. Some outside states have supplied arms or shielded the government in international forums, and resource interests such as jade and mining play a part in the political economy of the conflict.
• Congo (1996–2003 and after): Several neighbouring states and many armed groups fought over minerals, land, and political power. UN expert panels reported that minerals such as coltan and gold were traded through networks that financed armed groups, and that some neighbouring governments were involved. Religion played little role in most accounts; the wars were a contest over wealth and power.

Across these cases, the same pattern appears: local leaders use identity to recruit, while outsiders supply weapons, money, and cover. Stathis Kalyvas’s The Logic of Violence in Civil War (2006) shows that much violence is opportunistic, with personal grudges, looting, and property disputes settled under the label of a larger cause.

Human Selfishness, Named Honestly

Selfishness is not the trait of any faith, nation, or ethnic group. It is a universal human tendency, and it has several recurring forms in war:

• Greed: Control of diamonds, oil, coltan, or drug routes can make fighting profitable. Paul Collier and Anke Hoeffler’s “Greed and Grievance in Civil War” (2004) found that economic opportunity predicts civil war better than grievance alone.
• Power: Leaders gain legitimacy and survival by framing rivals as existential threats.
• Fear: When states weaken, people arm themselves for protection. Each group’s defence can look like aggression to its neighbours, a dynamic known as the security dilemma.
• Revenge and opportunism: Old disputes over land or debts are settled once the state’s authority has disappeared.
• Outside interest: Neighbouring governments, arms dealers, and resource companies often prefer a divided, weak country that they can influence.

Religion, ethnicity, and language are powerful tools for mobilising people who would otherwise hesitate. That is why they recur in war. They are rarely the first cause.

Colonial Rule Built the Categories

Colonial conquest was driven by economic extraction and state power. Colonial rule in the Americas centred on silver, land, and forced labour. In India, land-revenue systems and colonial census-taking helped fix caste and community categories that had once been more fluid (C. A. Bayly’s Caste, Society and Politics in India covers this). Colonial powers carved up Africa with little regard for existing societies.

In Rwanda, colonial authorities issued identity cards in the 1930s that recorded ethnicity, turning fluid social categories into fixed, hereditary ones. Mahmood Mamdani’s When Victims Become Killers (2001) argues that this colonial system hardened identities that later governments could exploit. The 1994 genocide, in which estimates commonly range from about 500,000 to 800,000 deaths, was organised by state and political actors. Many ordinary people participated under pressure and fear, and many others protected their neighbours. Individual responsibility matters, and so does the way institutions made killing possible.

Two Popular Explanations, Tested

Population pressure. Thomas Homer-Dixon’s Environment, Scarcity, and Violence (1999) argues that scarcity can raise conflict risk, especially where institutions are weak. But cross-national studies find that population growth alone predicts little about civil war. Many fast-growing countries are peaceful, and many violent conflicts occur where population pressure is modest. Scarcity matters most when governments are weak and elites can exploit it, which again points to selfishness and institutions.

Innate human nature. Richard Wrangham and Dale Peterson, in Demonic Males (1996), argue that coalitional aggression has deep evolutionary roots. Steven Pinker’s The Better Angels of Our Nature (2011) argues that violence has declined over long periods, which is hard to square with a fixed nature that always produces mass killing. Both views contain evidence, and the debate continues. Neither explains why the same populations can be peaceful in one decade and violent in the next. Those shifts track incentives, leadership, and institutions more closely than biology.

How AI and Robotics Can Counter Selfishness

Selfish actors rely on secrecy: hidden money, hidden arms shipments, false reports, and unaccountable forces. Technology can reduce that secrecy. It cannot change anyone’s motives, but it can make the costs of selfish choices more visible and harder to deny.

• Following the money and the minerals: Software can link mineral shipments to mines, traders, and refiners, supporting the due-diligence rules that the OECD and others have published. This targets the profit that makes some wars worth fighting.
• Exposing arms flows: Analysts already track weapons through shipment records, serial numbers, and imagery. Automated tools could scale this work and flag suspicious transfers to sanctions monitors.
• Verifying ceasefires and atrocities: Satellite imagery, open-source video, and drone surveys can document destroyed villages, troop movements, and mass graves. Amnesty International’s 2016 Decode Darfur project used volunteers to examine satellite images of burned villages.
• Early warning: Models that analyse news, economic shocks, and political events can flag rising tension. Uppsala University’s ViEWS project shows modest predictive accuracy, and false alarms are common. The value lies in prompting earlier human attention, not in automatic action.
• Accountability evidence: Well-organised, verified records help courts and commissions establish who gave orders and who profited.
• Clearing landmines: Robotic detectors and sensor-equipped drones can map contamination faster and more safely than manual survey. Clearance still needs funding and local knowledge.
• Humanitarian logistics: Drones have delivered blood and vaccines in Rwanda and Ghana. Aggregated, anonymised data can help agencies find displaced families and plan shelter and water.

The risks are just as serious, and selfish actors will exploit them:

• Surveillance misuse: Tools built to monitor conflict can be turned against civilians. Mass surveillance of minority communities, documented in several recent cases by human rights organisations and researchers, shows how data systems can target a group.
• Algorithmic bias: Models trained on biased records can turn old prejudice into automated suspicion.
• False alerts: A wrong warning can trigger a crackdown, a panicked flight, or a military response.
• Manipulated evidence: Synthetic images and video can fabricate atrocities or deny real ones, making verification harder.
• Profit-driven weapons: Autonomous weapons sold as products give selfish buyers and sellers a way to kill at lower political cost.
• Political cover: Governments may use technology to look responsive while leaving grievances and the selfish interests behind them untouched.

Institutions Must Make Selfishness Costly

Preventing mass violence means making selfish violence expensive and visible. That requires courts that hold leaders and profiteers accountable, sanctions that cut off illicit trade, a free press, fair access to land and resources, and outside governments that stop arming or shielding abusers. These are political choices made by people, and they can fail even where technology is abundant. In Rwanda, Bangladesh, Sri Lanka, and Myanmar, warnings were often visible; the failure lay in the response.

Technology is most useful when it helps institutions see clearly and act sooner. It is dangerous when it replaces their judgment or becomes another tool of the selfish. The aim is not a technical fix for human conflict, but better-informed human decisions that are harder to ignore.

The LIWARSE Lens

The 3 Absolute Laws of AI apply directly to this work:

• No Harm to Life: A system that labels a community as a threat is a potential instrument of harm. Designers should treat ethnic and religious targeting as a design failure and audit training data for bias before deployment.
• No Threat to Human Continuity: Autonomous systems should never make lethal decisions. Surveillance and targeting tools should be governed by international humanitarian law and meaningful human review.
• No Autonomous Self-Preservation: Early-warning and monitoring systems must not expand their own authority, data access, or resistance to correction. Humans must be able to inspect, override, and switch them off.

Shared identity is not the enemy. Selfishness, opportunity, and outside interests turn identity into a weapon. Safety of life depends on keeping that chain broken, with accountable humans at every link, and on building AI that exposes selfish conduct rather than serving it.

Sources for Further Reading

• Bayly, C. A. Caste, Society and Politics in India from the Eighteenth Century to the Modern Age. Cambridge University Press, 1999.
• Bose, Sarmila. Dead Reckoning: Memories of the 1971 Bangladesh War. Hurst, 2011.
• Collier, Paul, and Anke Hoeffler. “Greed and Grievance in Civil War.” Oxford Economic Papers 56 (2004).
• DeVotta, Neil. Blowback: Linguistic Nationalism, Institutional Decay, and Ethnic Conflict in Sri Lanka. Stanford University Press, 2004.
• English, Richard. Armed Struggle: The History of the IRA. Macmillan, 2003.
• Homer-Dixon, Thomas. Environment, Scarcity, and Violence. Princeton University Press, 1999.
• Kalyvas, Stathis N. The Logic of Violence in Civil War. Cambridge University Press, 2006.
• Mamdani, Mahmood. When Victims Become Killers: Colonialism, Nativism, and the Genocide in Rwanda. Princeton University Press, 2001.
• Pinker, Steven. The Better Angels of Our Nature. Viking, 2011.
• UN Group of Experts on the Democratic Republic of the Congo, reports to the Security Council (various years).
• UN Independent International Fact-Finding Mission on Myanmar. Report A/HRC/39/64, 2018.
• Wilson, Peter H. The Thirty Years War: Europe’s Tragedy. Harvard University Press, 2009.
• Wrangham, Richard, and Dale Peterson. Demonic Males: Apes and the Origins of Human Violence. Houghton Mifflin, 1996.
• UNHCR. Rohingya refugee population statistics, Bangladesh (current figures at unhcr.org).

Figures in this article are estimates drawn from the sources above and from commonly cited reporting. Readers should consult the originals before reusing any number.

— The LIWARSE Movement | liwarse.org
Safety of Life · Advancement of Life · Together