American criminal law has always recognized that the means used to commit an offense can matter almost as much as the offense itself. Robbery is serious; armed robbery is more serious. Assault is a crime; committing it with a deadly weapon can produce additional consequences. Society understands instinctively that when someone deliberately employs a tool that dramatically increases his ability to inflict harm, the law may reasonably take that increased capacity into account.
Artificial intelligence presents the same question on an unprecedented scale.
AI is not inherently dangerous. It is a tool capable of extraordinary good, from scientific discovery and medical research to education, engineering and economic productivity. But in the hands of a criminal, its extraordinary power becomes precisely the problem. AI can allow one individual to impersonate thousands of people, automate fraud, manufacture convincing evidence, clone voices, generate deepfakes, identify vulnerable targets and operate schemes at a scale that once required an organized criminal enterprise.
The law should recognize that amplification.
Consider the traditional confidence man. He might telephone ten elderly people during an afternoon and attempt to convince them that their grandchild is in trouble. Artificial intelligence potentially transforms that same criminal into a virtual fraud factory. Synthetic voices can impersonate relatives. Automated systems can conduct conversations. Publicly available information can personalize the deception. Generative technology can produce supporting documents, photographs and communications.
One criminal can potentially victimize thousands.
That distinction should matter at sentencing.
This does not mean that attaching AI to an ordinary crime should automatically produce some draconian mandatory sentence. Criminal law should preserve proportionality, intent and due process. Someone who unknowingly uses an AI-assisted service during an offense is different from someone who deliberately constructs an automated system designed to defraud 50,000 people.
The aggravating factor should be intentional weaponization.
If prosecutors can establish that an offender knowingly used artificial intelligence to increase the scale, sophistication, concealment or destructive capacity of a crime, enhanced penalties deserve serious consideration.
The principle is hardly radical. We already punish certain crimes more severely when aggravating circumstances increase their danger. Using sophisticated technology to magnify criminal capability should be viewed through a similar lens.
The strongest case involves impersonation and fraud.
Imagine criminals cloning the voice of a corporate executive and ordering an employee to transfer millions of dollars. Imagine fabricating a convincing video of someone’s daughter supposedly being held hostage. Imagine generating false evidence intended to destroy an innocent person’s reputation or interfere with a criminal investigation.
Those are not merely old-fashioned crimes performed on newer computers. AI changes the economics of deception by making extraordinarily convincing fabrication inexpensive, rapid and scalable.
There is another reason enhanced punishment deserves consideration: deterrence.
AI dramatically reduces the cost of attempting certain crimes. A fraudster who once needed employees, technical expertise and considerable time may increasingly be able to automate much of the operation. When the cost of committing an offense falls while the potential payoff increases, the expected consequences of deliberately abusing that technology become increasingly important.
Yet lawmakers should proceed carefully.
The worst response would be creating vague statutes criminalizing “harmful AI” without precisely defining prohibited conduct. Government has an unfortunate habit of responding to new technology by writing broad laws that eventually reach far beyond the conduct originally used to justify them.
AI itself must never become the crime.
Using AI to write anonymously, create satire, generate controversial political material, automate legitimate businesses, experiment with software or protect one’s privacy should not become suspicious merely because the technology is powerful. The First and Fourth Amendments do not become obsolete when computers become smarter.
The law should target criminal conduct, not technological capability.
A sensible framework would therefore focus on measurable aggravating factors: Did the defendant intentionally use AI to impersonate another person? Did AI enable the crime to operate at exceptional scale? Was synthetic media deliberately used to deceive victims? Did automation substantially increase financial losses or the number of victims? Was AI used to conceal the perpetrator’s identity or fabricate evidence?
Those questions examine conduct rather than criminalizing innovation.
Congress and state legislatures should also resist creating entirely separate criminal codes for artificial intelligence when existing statutes already address fraud, extortion, identity theft, stalking, forgery and other offenses. Where existing law works, prosecutors should use it. Where AI creates demonstrable gaps, lawmakers can establish carefully written enhancements.
America should remain a country in which technological innovation is presumed lawful. Entrepreneurs should not need government permission to experiment with powerful tools.
But liberty requires accountability.
Artificial intelligence gives individuals capabilities once reserved for corporations, intelligence agencies and sophisticated criminal organizations. That democratization of power will produce extraordinary benefits, but it also means a malicious individual can potentially cause extraordinary damage.
The law must recognize both realities.
AI should remain widely accessible, innovative and free from unnecessary government control. But when someone deliberately turns that extraordinary capability into an instrument of fraud, extortion, impersonation or other serious crime, society is justified in treating the amplification itself as an aggravating factor.
Greater technological power should bring greater personal responsibility.
And when that power is intentionally weaponized against innocent people, the punishment should reflect not merely the crime committed, but the extraordinary machinery deliberately placed behind it.