1. When the Chips Are Down
“When the chips are down, these… these civilized people, they’ll eat each other.”
— The Joker, The Dark Knight
The Joker’s argument is more interesting than the simple claim that people are secretly evil. He is testing something deeper: how much of civilization’s morality survives when the conditions supporting civilization are removed?
When security disappears, survival becomes more important. When trust collapses, cooperation becomes dangerous. When the rules stop protecting people, people may stop obeying the rules. The Joker believes that morality is therefore conditional: it works while the surrounding system works.
That was a fictional experiment. But artificial intelligence may create a real one. For thousands of years, human societies have depended upon human beings. We needed farmers, workers, soldiers, merchants, engineers, teachers, administrators and consumers. Even the most powerful person ultimately depended upon millions of other people.
AI and robotics could begin changing that dependency. What happens when the powerful no longer need the majority of people in the same way? That is the Joker Test.
2. Why Humans Needed Morality
Human morality did not appear in a vacuum. Humans are social animals, and survival increasingly depended upon cooperation.
A group in which everyone steals from everyone else, breaks every promise and responds to every disagreement with violence cannot easily build a civilization. Trust becomes impossible. Trade becomes difficult. Families become unstable. Long-term projects become dangerous.
Rules therefore emerged. Some were practical: don’t steal, don’t kill members of the group, keep promises, protect children, respect property, reciprocate cooperation. Others became moral principles: this is right; that is wrong.
Religion strengthened many of these systems by placing social rules inside a larger story. Rules could be presented as sacred rather than merely convenient, with reward and punishment extending beyond ordinary human enforcement.
Perhaps morality became powerful partly because society needed it to function. The rules were not merely describing an ideal human being. They were helping create a society in which millions of imperfect human beings could live together.
3. Morality as Civilization’s Anesthesia
There is an uncomfortable way to look at this. Perhaps morality is partly civilization’s psychological anesthesia.
Human beings have primitive impulses: aggression, fear, greed, revenge, status seeking and the desire to take what others possess. Civilization cannot simply eliminate these impulses. Instead, it creates systems that restrain and redirect them.
If I refrain from stealing only because I am afraid of being caught, my behavior depends on enforcement. If I believe stealing is morally wrong, the restraint has moved inside my mind.
Religion, morality, law and culture therefore perform overlapping functions. They make cooperation psychologically easier by turning social necessities into internal convictions.
Calling morality a “placebo” does not necessarily mean it is fake. A placebo can produce a real effect. The deeper question is whether the moral system remains equally powerful when the conditions that produced the need for it begin to change.
4. The Hidden Foundation: Mutual Dependence
For most of human history, power existed inside a web of dependence.
A king needed farmers to produce food and soldiers to defend his territory. He needed administrators to collect taxes. Merchants needed customers. Employers needed workers. Workers needed employers. Governments needed taxpayers. Companies needed consumers.
Even enormous concentrations of wealth did not eliminate these dependencies. The powerful could command more resources, but they could not simply replace the entire human system underneath them.
This created an important constraint on power. If you depend upon other people, you have reasons to keep the relationship functioning. Those reasons do not necessarily come from kindness. They can come from necessity.
The social contract therefore has an economic foundation that is easy to overlook. People cooperate partly because they need one another. This creates a profound connection between economics and ethics.
5. When Technology Removes Dependence
Now consider what happens when technology begins removing those dependencies.
AI can perform cognitive tasks. Robots can perform physical tasks. Autonomous systems can increasingly operate without continuous human control. Advanced military technology can reduce the need for mass armies in some forms of warfare, while automated systems can reduce the need for large numbers of human operators in some forms of administration and security.
None of this means that humans will suddenly become unnecessary. That would be an extreme claim. But the direction matters.
For the first time, technology is potentially capable of replacing not only human muscles, but increasingly human cognition as well.
That changes the equation. The historical ruling structure was: power depends on people. The emerging possibility is: power controls machines that perform much of what people used to provide. That is a very different civilization.
6. The Rise of the Independent Power Structure
Imagine a system in which a relatively small amount of capital controls enormous productive capacity.
The system owns or controls computing infrastructure, energy, robotics, factories, networks, intellectual property and increasingly capable AI. The machines produce more machines. AI designs improvements. Robots maintain infrastructure. Automated systems manage logistics. Capital finances further expansion.
The system becomes increasingly capable of reproducing and expanding itself with relatively little additional human labor.
This is a thought experiment, not a prediction that such a system will inevitably emerge exactly this way. But it exposes the central problem: what happens to political and social power when economic power becomes much less dependent on mass human participation?
For centuries, ordinary people possessed an indirect source of power simply because society needed them. They worked, produced, consumed, paid taxes, fought wars, organized, struck and could collectively refuse to cooperate. If technology progressively eliminates those forms of dependence, something much more fundamental than employment changes. The bargaining structure of civilization changes.
7. The Bargaining Power Problem
A worker’s bargaining power is not determined only by how much money the worker has. It also depends upon how difficult the worker is to replace.
A highly replaceable worker has little leverage. A worker possessing scarce skills has more. A large group of workers who cannot easily be replaced can have enormous collective power.
This is why labor has historically mattered politically as well as economically.
But suppose AI makes large categories of cognitive work replaceable, robotics makes large categories of physical work replaceable, and automated systems reduce the need for human administrative, security and operational roles.
The question is no longer simply how many jobs will disappear. The deeper question is how much bargaining power disappears when human beings become less economically necessary.
A society can theoretically provide people with income after their labor becomes unnecessary. But income provided by the system is different from power over the system. That distinction may become one of the defining questions of the AGI era.
This connects directly to the earlier article Will AGI Let Elites Rule Forever—and Make Human Bargaining Power Disappear?, which examines the same problem from the perspective of power and ownership.
8. The Joker Test Meets AI
This brings us back to the Joker. His experiment asks what happens when the incentives and conditions supporting civilized behavior are disrupted.
AI creates a different experiment. What happens when economic necessity itself is disrupted?
A company replaces ten workers with an AI system. The decision may be entirely rational. Costs fall. Productivity rises. The company becomes more competitive. Investors reward it. Competitors imitate it. Eventually, refusing to automate may itself become economically irresponsible.
No villain is required. That is what makes the problem difficult.
One company automates because it must compete. Another follows. Workers become easier to replace. Training pipelines shrink. Human expertise becomes scarcer. Capital becomes more productive. Ownership becomes more valuable.
If efficiency rewards the removal of human dependence, how long can moral restraints survive against that incentive?
AI does not have to become immoral. It may simply become more efficient than morality is willing to be.
9. Nature Has No Moral Tribunal
There is an even older version of this problem in nature.
A lion eats a deer. A snake eats a mouse. The deer is not guilty. The mouse has committed no crime. But nature does not convene a moral tribunal. There is no question of whether the prey deserves to survive.
Natural selection operates without concepts such as justice, innocence or fairness. Humans are unusual because we have tried to place moral constraints around these forces.
We tell ourselves that the strong should not simply consume the weak. We create laws. We protect children. We care for the sick. We help people who cannot compete. We restrain our ability to use power simply because we possess it.
Nature explains what can happen. It does not tell us what humans should permit.
That distinction may become crucial in an AI civilization. Technology could make our productive systems increasingly resemble nature in one particular respect: relentless optimization without moral preference.
10. When Efficiency Becomes the Strongest Force
Imagine an executive choosing between two production systems. One employs 10,000 people. The other employs 1,000 people and uses advanced AI and robotics. The second system produces more at lower cost.
If the executive chooses the first system purely to preserve employment, a competitor may choose the second. The competitor gains an advantage. More firms follow. Soon the ethical choice becomes economically difficult to sustain.
This is the deeper danger of technological efficiency. It does not need to arrive as a command. It can arrive as a competitive pressure.
Every individual decision can be rational while the collective result is transformative. The danger is therefore not necessarily that someone decides humans no longer matter. It may be that everyone discovers that treating humans as economically necessary is becoming increasingly expensive.
At that point, efficiency begins to function almost like a force of nature. The question becomes whether civilization is willing to place limits on it.
11. Can Ethics Survive Without Necessity?
This may be the hardest question.
There is another assumption hidden inside the idea of an AI-abundant future: that abundance will automatically be shared. But technology changes what can be produced; it does not, by itself, change who owns the productive system.
There is also a cultural problem. In a capitalist system, ownership is normally understood as an entitlement earned through investment, risk and work. A person who becomes extraordinarily wealthy can sincerely believe, “I earned this through my hard work,” even when the productive power of AI and capital has become enormous. Philanthropy may happen, but it is voluntary; it cannot be assumed simply because abundance exists.
Suppose society reaches a point where machines can produce most necessities, maintain infrastructure and perform much of the work previously performed by humans. Would we still believe that every person deserves meaningful economic participation?
Probably many people would. But believing something and possessing the power to enforce it are different things.
Historically, ethics operated inside a world of mutual dependence. The farmer needed the soldier. The ruler needed the taxpayer. The employer needed the worker. The worker needed the employer. The consumer needed the producer. Everyone had something the other side needed.
AGI could weaken some of those relationships. Human beings could become materially less necessary at exactly the moment when technology makes the productive system vastly more powerful.
The moral question would then become: if people are no longer necessary for production, do they remain entitled to a meaningful share of production?
That is not an engineering question. It is a civilization question.
This is also where the article Universal Basic Beggars: The Hidden Risk of an AI-Abundant World becomes relevant: material security may rise even as dependence, ownership and bargaining power change.
12. The Real Test of Civilization
The easiest time to be ethical is when ethics and self-interest point in the same direction. The difficult test comes when they diverge.
A society can proclaim human dignity when human beings are indispensable. The harder question is what happens when they are not.
The Joker’s original wager was that civilized behavior might collapse when the conditions supporting it disappear. The AI version is more subtle. Perhaps civilization will not collapse into chaos. Perhaps it will become extremely efficient, extremely productive and extremely unequal in power.
Perhaps everyone will still speak about dignity, fairness and human rights. But underneath the language, the dependency structure may have changed.
And that is where the real Joker Test begins.
Can humans preserve moral constraints when technology gives some humans unprecedented freedom from dependence on everyone else?
Because the ultimate danger of AGI may not be that machines become evil. It may be that efficiency becomes so powerful that morality becomes optional for those who control the machines.
Perhaps that is the final lesson of the Joker: Civilization is not tested when doing the right thing is easy. It is tested when doing the right thing becomes costly.
Related reading