Artificial intelligence has become one of the defining technologies of the twenty-first century. It may transform medicine, education, science, transportation, manufacturing, finance, communication, and warfare. It may also change how people work, how governments exercise power, and how companies create wealth.
Therefore, the race for leadership in artificial intelligence carries enormous economic and political importance. The United States wants to protect its technological advantage. China wants to reduce its dependence on American companies and build systems of its own. Both countries treat advanced AI as a strategic asset, not merely as another commercial product.
Unsurprisingly, this rivalry has produced accusations, restrictions, sanctions, and suspicion. American politicians frequently accuse Chinese companies and state institutions of stealing intellectual property. They point to cyberespionage, industrial espionage, forced technology transfers, and the copying of foreign products. Chinese officials, meanwhile, argue that American restrictions aim to preserve a technological monopoly and prevent China from competing fairly.
Most discussions stop at this point. Theft is wrong. Innovation is good. Intellectual property deserves protection. Therefore, anyone who copies technology must harm progress.
However, the issue may not be so simple.
Throughout history, technological imitation has often accelerated development. Countries that copied foreign machines, production methods, and scientific techniques later became major innovators themselves. Companies also learned from rivals, improved existing inventions, and created products that reached millions of people.
Consequently, the uncomfortable question deserves serious attention. Could humanity sometimes benefit when knowledge spreads faster than governments and corporations would prefer? Could imitation, reverse engineering, or even illegal technological theft occasionally accelerate progress?
I may be wrong. Nevertheless, I believe the answer could sometimes be yes.
Innovation has always depended on imitation
People often imagine innovation as the work of a solitary genius. According to this romantic story, an inventor enters a laboratory, has a brilliant idea, and creates something entirely new. Society then rewards that person for an original contribution.
Reality looks very different.
Almost every important invention rests on thousands of earlier discoveries. Engineers improve existing machines. Scientists build on previous experiments. Software developers reuse ideas, methods, and code created by others. Even revolutionary products usually combine technologies that already existed.
Thomas Edison did not create electric lighting from nothing. He improved earlier designs and developed a commercially practical system around them. The Wright brothers studied experiments conducted by previous aviation pioneers. Modern smartphones combine telecommunications, computing, cameras, satellites, batteries, touchscreens, and countless other technologies developed over many decades.
Artificial intelligence follows the same pattern. Modern AI systems rely on mathematical research, computer science, neuroscience, statistics, linguistics, and enormous public investment. Universities published many of the ideas that now power commercial models. Governments funded research laboratories. Open-source communities created essential software. Private companies then combined these resources with enormous computing power and vast amounts of data.
Learning from the past
Therefore, no company can honestly claim that it created modern artificial intelligence alone. Every major AI laboratory stands on a foundation built by previous generations of researchers.
Countries also learn through imitation. In fact, many of today’s richest states once copied technologies from foreign rivals.
During the nineteenth century, the United States borrowed heavily from European industrial methods. American manufacturers studied British machinery and production systems. Britain often tried to prevent skilled workers and technical knowledge from leaving the country, yet those efforts did not stop the spread of industrial expertise.
Later, Japan studied Western manufacturing, military organization, engineering, and education. After the Second World War, Japanese companies became famous for adapting and improving American and European products. At first, critics often dismissed Japanese goods as cheap copies. Over time, however, Japanese manufacturers became global leaders in automobiles, electronics, robotics, and industrial production.
South Korea followed a similar path. Its companies learned from Japanese and American firms. They entered existing industries, copied successful methods, improved quality, and gradually developed technologies of their own. Today, South Korean companies compete at the highest level in electronics, semiconductors, automobiles, shipbuilding, and telecommunications.
China has followed the same broad pattern on a much larger scale. It began as a manufacturing base for foreign companies. Chinese firms learned how global supply chains worked. They copied products, hired foreign-trained specialists, reverse-engineered machines, and invested heavily in domestic research. Some companies continued to imitate. Others became genuine innovators.
This historical pattern does not prove that theft is morally right. However, it does show that imitation often comes before innovation. Yesterday’s copier can become tomorrow’s inventor.
The difference between learning, copying, and stealing
Before defending any form of technological imitation, we must distinguish several very different activities.
Learning from a competitor is not theft. Reading a scientific paper is not theft. Hiring an engineer who brings general experience from a previous job is not theft. Reverse-engineering a legally purchased product may also be lawful in some circumstances.
Likewise, building a similar product does not automatically mean that someone stole it. Companies often reach similar solutions because they face the same technical problems. Engineers can also improve a public idea without accessing confidential information.
At the same time, real theft clearly occurs. Hackers can break into corporate networks and steal source code. Employees can copy trade secrets before joining a rival. Intelligence services can target research laboratories. Companies can violate patents or use confidential designs without permission.
These actions differ morally and legally. Therefore, an honest discussion cannot place all forms of imitation into one category.
Nevertheless, the boundaries often become unclear. Scientific knowledge spreads through publications, conferences, employee movement, open-source projects, and informal communication. A company may call something a trade secret even when many researchers already understand the basic principle. Governments may also accuse rivals of theft when they really fear competition.
As a result, political language often hides more than it explains. When American officials accuse China of stealing technology, some allegations may involve genuine espionage. Others may describe aggressive but legal learning. Still others may reflect anxiety that China has become capable of competing.
The AI race changes the incentives
Artificial intelligence develops unusually quickly because every improvement helps produce the next improvement.
A stronger model can assist programmers, summarize research, generate training data, test software, and help scientists explore new ideas. Better chips make larger models possible. Better algorithms reduce computing costs. New tools allow companies to train systems more efficiently.
Therefore, progress accumulates.
This creates an unusual economic problem. One company may spend billions of dollars training a frontier model. It may hire the best researchers, purchase advanced chips, build data centers, and experiment for years. Then competitors may study the resulting system and learn from its visible behavior.
They may not steal the code. They may not access private files. However, they can still discover useful techniques through published research, employee movement, model testing, and open-source replication.
Consequently, the first company bears much of the cost, while later competitors reduce their own costs by learning from the pioneer.
For the original developer, this creates a serious problem. Its rivals can close the gap more quickly. Investors may lose confidence. The company may struggle to recover the money it spent.
For society, however, the outcome may look different. Competition can lower prices. New companies can challenge dominant firms. Researchers can improve the original idea. Consumers can access better tools sooner.
The interests of one company and the interests of humanity do not always align.
A corporation naturally wants to preserve its advantage for as long as possible. Society, meanwhile, may benefit when useful knowledge spreads widely. This tension lies at the center of the debate.
Why competition usually benefits consumers
Economic history repeatedly shows that monopolies often slow progress.
A dominant company may still innovate. However, it faces less pressure to reduce prices, improve quality, or respond quickly to users. When customers have no realistic alternative, the company can become complacent.
Competition changes these incentives.
When one AI company releases a stronger model, rivals must react. They improve their systems, reduce prices, add features, or enter new markets. When Chinese laboratories create efficient models, American companies face pressure to cut costs. When American firms develop new capabilities, Chinese companies invest more aggressively.
Consumers benefit from this rivalry.
Models improve more quickly. Prices fall. Smaller businesses gain access to powerful tools. Researchers receive new systems for scientific work. Developers can choose among several providers instead of depending on one company.
Competition also reduces political risk. No single company should control a technology that may influence education, employment, communication, surveillance, and access to information. If one corporation dominates advanced AI, it may gain extraordinary power over governments and citizens.
International competition can prevent that outcome. Even hostile rivalry may limit monopoly power.
Of course, competition can also create dangers. Companies may release unsafe systems too quickly. Governments may weaken regulation because they fear losing the race. Researchers may prioritize speed over caution.
Nevertheless, the absence of competition could create even greater dangers. A global AI monopoly would possess enormous economic, cultural, and political influence. Therefore, rivalry between the United States and China may provide some benefits, even when the relationship remains tense.
Could technological theft increase overall welfare?
This is where the argument becomes uncomfortable.
Imagine that one country illegally obtains valuable AI techniques from another. The original developer loses part of its advantage. It may lose customers, revenue, and investor confidence.
Normally, we would describe this outcome as harmful.
However, suppose that the copied knowledge allows several companies to build cheaper and more capable AI systems. Doctors use those systems to improve diagnoses. Researchers use them to develop medicines. Teachers use them to provide individual instruction. Engineers use them to design cleaner energy systems. Factories use them to reduce waste.
In that case, millions or billions of people could benefit.
From a strictly utilitarian perspective, the total gains might exceed the losses suffered by the original company. A corporation may lose billions of dollars, while society gains medical advances, higher productivity, cheaper services, and faster scientific progress.
That calculation does not automatically make theft morally acceptable. Rights, laws, fairness, and trust still matter. Furthermore, we cannot simply take anything from anyone whenever we predict a larger social benefit.
Still, the possibility reveals an important truth. The consequences of technological theft can differ from the consequences of ordinary theft.
If someone steals a car, the owner loses the car. Only one person can use it at a time. Knowledge works differently. When someone copies an algorithm, the original owner still possesses it. The thief gains a copy, but the original does not disappear.
Economists call knowledge non-rivalrous because many people can use the same idea simultaneously. This feature makes intellectual property fundamentally different from physical property.
The law creates temporary exclusivity because innovators need rewards. Yet society also benefits when knowledge spreads. Therefore, intellectual property always involves a compromise between private incentives and public access.
Why intellectual property still matters
A provocative argument for imitation must not ignore the strongest argument on the other side.
Developing frontier AI systems costs enormous amounts of money. Companies must purchase advanced chips, build data centers, hire highly skilled researchers, obtain electricity, collect data, and conduct repeated experiments. Many projects fail. Even successful systems may not immediately produce profits.
Investors accept these risks because they expect future returns.
If competitors could immediately copy every breakthrough without paying anything, companies might reduce their investments. Why spend billions of dollars on research if a rival can take the results the next day?
Innovation could then slow down.
Patents, copyrights, and trade-secret laws try to solve this problem. They give creators a period of protection. During that time, innovators can recover some of their costs. Afterward, knowledge may spread more widely.
This system never works perfectly. Companies can abuse patents. Copyright can last too long. Large corporations can use legal threats to suppress smaller competitors. Patent trolls can demand money without creating useful products.
Nevertheless, some protection remains necessary.
Without it, private companies may avoid expensive research and focus only on projects that produce immediate profits. Society would then lose inventions that require years of investment.
Therefore, the real question is not whether intellectual property should exist. It should. The real question concerns how much protection society should grant and for how long.
Too little protection can weaken innovation. Too much protection can create monopolies and slow the spread of knowledge.
Monopoly can also harm innovation
Defenders of strict intellectual-property protection often focus on the danger of copying. However, they sometimes ignore the danger of excessive control.
A company that owns a crucial technology can charge high prices. It can block competitors. It can limit access. It can decide which countries, universities, or businesses may use the technology.
This control may benefit shareholders. It may not benefit humanity.
Medical patents provide a clear example. Drug companies need profits because pharmaceutical research costs a great deal. Yet extremely high prices can prevent poor patients from accessing life-saving treatments. Society must therefore balance innovation incentives against human needs.
Artificial intelligence may create a similar conflict.
Suppose a small number of American companies control the most powerful AI systems. They could decide who receives access, how much users pay, and what information the systems provide. They could also cooperate with intelligence agencies or shape political debate.
In that situation, Chinese competition might provide a useful counterweight. Even if Chinese companies copied some ideas, their presence could reduce prices and prevent a complete American monopoly.
The reverse also applies. If Chinese companies dominated AI, American and European competition would remain essential.
Therefore, international rivalry can protect consumers from concentrated power. It can also prevent one political system from controlling the global development of artificial intelligence.
China does not only copy
Western discussions often portray China as a country that steals ideas but rarely creates anything original. This view no longer reflects reality.
China still faces accusations of industrial espionage, patent violations, and forced technology transfers. Some allegations appear credible. However, Chinese companies and universities also conduct genuine research. They publish scientific papers, build advanced products, and employ highly skilled engineers.
China leads or competes strongly in several technological sectors. These include batteries, electric vehicles, solar panels, telecommunications, drones, manufacturing, and some areas of artificial intelligence.
Moreover, copying and innovation can exist within the same economy. A company may begin by imitating foreign products and later develop original technologies. Japan and South Korea followed similar paths.
Therefore, dismissing every Chinese achievement as theft creates a false sense of security in the West. It encourages American and European politicians to blame China rather than confront their own weaknesses.
The United States still possesses enormous advantages. It has leading universities, powerful technology companies, deep capital markets, advanced chip designers, and a strong culture of entrepreneurship. However, it cannot preserve leadership through accusations alone.
It must continue to invest, educate, compete, and innovate.
The United States also benefited from foreign knowledge
The United States often presents itself as the natural home of invention. In reality, it also gained enormously from foreign knowledge, foreign scientists, and foreign capital.
European immigrants brought technical skills during American industrialization. Scientists who fled fascism helped transform American universities and military research. After the Second World War, the United States recruited German engineers and scientists, including people connected to the Nazi regime.
American companies also studied Japanese manufacturing methods. Silicon Valley benefited from researchers born and educated around the world. Many leading American AI scientists came from Europe, China, India, Canada, and other regions.
Therefore, American technological success never developed in national isolation. It depended on global knowledge flows.
This does not equal theft. However, it weakens the simplistic claim that innovation belongs entirely to one nation.
Knowledge moves across borders. Scientists collaborate. Employees migrate. Ideas spread. Governments can slow this process, but they cannot completely stop it.
Export controls and their unintended effects
The United States has tried to limit China’s access to advanced semiconductors and chip-making equipment. The goal appears clear. Washington wants to slow Chinese military and AI development.
In the short term, these restrictions may work. Advanced AI systems require powerful chips, and the United States controls important parts of the semiconductor supply chain.
However, restrictions can also create unintended consequences.
When a country loses access to foreign technology, it gains a stronger incentive to develop domestic alternatives. China may invest more heavily in chips, manufacturing tools, software, and research because it can no longer rely on American suppliers.
As a result, export controls may delay China while also encouraging it to become more independent.
History offers many examples of sanctions producing this effect. Restrictions create pain, but they also stimulate substitution. Countries search for new suppliers, build local industries, and redesign products around available materials.
Therefore, the United States faces a difficult choice. If it shares advanced technology freely, China may close the gap faster. If it blocks access, China may accelerate its efforts to achieve technological independence.
Neither strategy guarantees permanent American dominance.
The security problem changes everything
Artificial intelligence is not merely a commercial technology. It also affects national security.
Advanced AI may assist cyberattacks, intelligence analysis, military planning, surveillance, propaganda, autonomous weapons, and drone operations. It may help governments identify targets, break into computer systems, or monitor entire populations.
Therefore, an argument that supports broad knowledge sharing in civilian science becomes less persuasive when military applications enter the picture.
A copied medical algorithm may save lives. A copied targeting system may help kill people. A model that improves industrial efficiency may benefit society. A model that assists authoritarian surveillance may strengthen repression.
Consequently, governments cannot treat every AI technology in the same way.
Some knowledge should remain open. Some research deserves temporary commercial protection. Some military systems require strict security.
The challenge lies in drawing these boundaries.
Governments often use national security as an excuse for economic protectionism. Companies may label ordinary commercial technology as strategically sensitive because they want to block competitors. At the same time, genuine security threats exist.
Therefore, policymakers must distinguish carefully between protecting citizens and protecting corporate profits.
Open source complicates the entire debate
Artificial intelligence differs from many older technologies because a large part of its development occurs openly.
Researchers publish papers. Universities share discoveries. Developers release software libraries. Open-source communities create models that anyone can download, examine, and improve.
Companies also contribute to open-source projects because they benefit from shared standards and tools. Even fierce competitors rely on the same programming languages, scientific papers, and software frameworks.
This makes the line between cooperation and competition increasingly difficult to identify.
A Chinese laboratory may build on an American research paper. An American company may use an open-source tool created by European or Chinese developers. Researchers may study a rival’s model and publish improvements that everyone can use.
None of this necessarily involves theft.
In fact, open collaboration has driven much of AI’s rapid development. Scientists gain prestige by publishing discoveries. Companies attract talent by sharing research. Developers improve tools collectively.
Therefore, the AI revolution already depends on controlled forms of knowledge sharing.
The strongest argument for openness says that humanity faces common problems. Climate change, disease, aging, poverty, and energy shortages affect people across national borders. If AI can help solve these problems, restricting every breakthrough may slow solutions that the world urgently needs.
The strongest argument against openness says that dangerous capabilities can also spread. Criminals, dictatorships, and terrorist groups may use advanced systems. Open-source AI may reduce the ability of governments to control powerful technologies.
Again, no simple answer exists.
Competition can accelerate safety as well as capability
Many people assume that competition only encourages companies to build more powerful systems. However, competition can also improve safety.
A company that offers a more reliable model can attract customers. Businesses do not want systems that invent facts, leak confidential data, or behave unpredictably. Governments also prefer tools that they can audit and control.
Therefore, companies may compete on security, transparency, accuracy, and efficiency.
Chinese and American researchers can also expose each other’s weaknesses. Rival laboratories may discover vulnerabilities that the original developers missed. Independent testing can reveal manipulation, bias, or dangerous capabilities.
At the same time, competitive pressure can clearly weaken safety. Companies may rush products to market. Governments may suppress warnings because they fear falling behind. Researchers may treat caution as a luxury.
The outcome depends on regulation and incentives.
Competition alone will not guarantee safety. Yet monopoly will not guarantee it either. A dominant company may conceal problems, resist oversight, and define safety according to its own interests.
Therefore, society needs both competition and strong public rules.
The moral problem of rewarding thieves
Even when technological theft produces social benefits, serious moral objections remain.
A company that invests money and effort deserves some reward. Employees deserve protection against the theft of their work. Businesses need trust to cooperate. International agreements lose meaning when governments ignore them.
If society openly celebrates theft, it may undermine these principles.
Moreover, powerful states would probably benefit more than weak ones. Intelligence agencies and large corporations possess the resources to steal valuable technology. Small businesses and poorer countries cannot defend themselves as effectively.
Therefore, a world without intellectual-property rules might not become fairer. It might simply reward the strongest thieves.
This creates a distinction between understanding the positive consequences of diffusion and endorsing theft as a general moral rule.
We can recognize that stolen knowledge sometimes benefits society without arguing that theft should become legal or admirable. Consequences matter, but so do institutions, trust, and fairness.
The utilitarian dilemma
From a utilitarian perspective, the issue remains particularly difficult.
Utilitarianism asks which action produces the greatest overall well-being. If copying an AI breakthrough harms one company but helps billions of people, the calculation may favor diffusion.
However, utilitarian reasoning must include long-term consequences.
If repeated theft destroys incentives for research, future innovation may decline. If espionage increases international hostility, the risk of conflict may rise. If companies stop cooperating across borders, scientific progress may slow.
Therefore, we cannot judge one act in isolation.
A single stolen technology might produce enormous benefits. A global system that normalizes theft might produce instability, mistrust, and lower investment.
The best policy may therefore combine strong but limited protection with faster public access.
Governments could shorten some monopoly rights. They could require licensing under fair conditions. They could fund open research. They could reward inventors directly while allowing broader use of important technologies.
Such policies would spread knowledge without depending on espionage.
Better alternatives to technological theft
If society benefits from rapid diffusion, it should create legal ways to achieve it.
Public research provides one option. Governments can fund AI projects and require that researchers publish the results. Universities can develop open tools. International organizations can support shared scientific infrastructure.
Compulsory licensing provides another option. Under this system, governments allow others to use protected technology while requiring payment to the inventor. This approach already appears in debates about essential medicines.
Patent pools could also help. Companies could share important technologies under common rules. Smaller firms would gain access, while inventors would still receive compensation.
Prizes offer another possibility. Governments or foundations could reward major breakthroughs and then release them publicly. The creator would receive money, while society would receive open access.
These approaches may produce many of the benefits associated with imitation without rewarding illegal conduct.
However, governments rarely act quickly enough. Powerful companies lobby to preserve exclusive rights. National rivalries block international cooperation. Therefore, knowledge often spreads through informal, controversial, or illegal channels instead.
Europe risks becoming a spectator
The AI rivalry usually focuses on the United States and China. Europe often appears only as a regulator or consumer.
This creates a serious problem.
Europe possesses strong universities, skilled researchers, advanced industries, and wealthy economies. However, it often lacks the scale, investment, and risk tolerance needed to build globally dominant technology companies.
As a result, Europeans may end up choosing between American and Chinese systems.
That would weaken European independence. It could also give foreign companies influence over European data, infrastructure, employment, and public debate.
Therefore, Europe should not merely criticize American monopolies or Chinese copying. It should build its own AI capacity.
European governments need larger research programs, better access to capital, stronger computing infrastructure, and more ambitious industrial policies. At the same time, Europe should avoid regulation that protects established companies from new competitors.
A third center of AI power would benefit the world. It would reduce dependence on both Washington and Beijing. It would also increase competition and provide alternative standards.
Cooperation remains necessary
Despite their rivalry, the United States and China cannot avoid cooperation entirely.
Both countries face common risks from artificial intelligence. Advanced systems may create cyber threats, misinformation, unemployment, military instability, and loss of human control.
No country can solve these problems alone.
If American and Chinese laboratories race without communication, they may misinterpret each other’s actions. Governments may assume that every technical advance serves military purposes. This distrust could increase the risk of conflict.
Therefore, the two powers need at least limited agreements. They could discuss military uses of AI, autonomous weapons, nuclear command systems, cyberattacks, and safety testing.
Competition and cooperation can exist at the same time.
The United States and the Soviet Union competed fiercely during the Cold War. Yet they still negotiated arms-control agreements because both understood the danger of uncontrolled escalation.
AI may require a similar approach. Nations can compete economically while cooperating to prevent catastrophe.
My opinion
I may be wrong. However, I suspect that history will judge technological copying in the AI era more positively than many politicians expect.
Humanity usually benefits when powerful organizations compete rather than dominate. Competition lowers prices, accelerates innovation, and limits monopoly power. It also gives consumers and governments alternatives.
Technological imitation often strengthens this competition. A country copies, learns, improves, and eventually begins to innovate. History shows this pattern repeatedly.
Therefore, I cannot accept the simple claim that every copied idea harms humanity.
Sometimes the original developer loses while society gains. A company may lose market share, yet millions of people receive cheaper technology. A country may lose a temporary advantage, yet global scientific progress accelerates.
That does not mean we should celebrate all theft.
Cyberattacks remain dangerous. Stealing personal data violates individual rights. Military espionage increases international tension. Companies also deserve a reasonable opportunity to recover the money they invest in research.
Nevertheless, absolute protection can become harmful as well. It can create monopolies, block poorer countries, raise prices, and slow the spread of useful knowledge.
The real objective should therefore involve balance.
Society must reward innovators without allowing them to control essential knowledge forever. Governments must protect national security without using it as an excuse for permanent economic domination. Companies must compete, but they must also follow safety rules.
Most importantly, humanity should not confuse corporate interests with human interests.
A company wants exclusive control because exclusivity creates profit. A government wants technological dominance because dominance creates power. Humanity, however, benefits from broad access, rapid progress, strong competition, and peaceful cooperation.
These goals often conflict.
Turning theft into virtue
Perhaps technological theft sometimes produces good consequences. That possibility does not turn theft into a universal virtue. However, it should force us to question simple moral slogans.
Innovation has never respected borders completely. Ideas travel. Scientists move. Engineers copy. Companies improve. Countries learn from one another, whether governments approve or not.
The United States copied before it led. Japan copied before it led. South Korea copied before it led. China copied and now increasingly innovates.
Therefore, the future of artificial intelligence will not belong permanently to one country or one company. Knowledge will spread. Rivals will imitate one another. New competitors will emerge.
The important question is not whether we can stop this process entirely. We probably cannot.
The important question is whether we can shape it so that competition accelerates human progress without destroying trust, security, and the incentives needed to create the next breakthrough.
That balance may determine whether artificial intelligence becomes another tool of monopoly and national domination or a technology that improves life for humanity as a whole.
Leave a Reply