Can the law defeat AI in the service of crime?

We live in an era in which the boundaries between physical and digital identity have become so blurred that the law, in its traditional form, is starting to lose its footing. As a lawyer, I am witnessing a new kind of conflict: on one side is the judiciary, slow and procedural by nature, and on the other is high-tech crime powered by artificial intelligence, which knows no borders, no working hours and no moral compass.

The question is no longer whether technology will be misused, but how we will defend ourselves when the metaphor of “identity theft” turns into the literal digital abduction of your likeness and your actions.

Deepfake: the perfect crime without fingerprints?

Deepfake technology has travelled from entertaining social-media filters to the most dangerous tool of social engineering. In legal terms, we are facing a phenomenon that goes beyond classic defamation.

Imagine the scenario: the chief executive of a large company receives a video call from the owner, ordering an urgent transfer of millions to an unknown account. The voice is identical, the facial expressions match the urgency of the situation. Only after the transaction is it revealed that the “owner” was an AI algorithm.

Where is the problem for a lawyer? The Serbian Criminal Code recognises fraud (Article 208), but proving the intent of someone who used anonymous AI servers in a jurisdiction that does not cooperate through Interpol is a legal nightmare. We currently have no specific offence of “digital impersonation for the purpose of fraud”, which leaves room for the defence to manipulate evidence.

Generative AI: a factory of fake evidence and the “liar’s dividend”

One of the most chilling aspects of AI in the justice system is the ability to create fake evidence that passes forensic examination. AI can generate:

Audio recordings of conversations that never took place.

Documents with perfectly replicated signatures and stamps.

Images that “prove” a person's presence at a certain place.

One of the most chilling aspects of AI in the justice system is the ability to create fake evidence that passes forensic examination.

Slaviša Cincarević

On the other hand, there is the “liar’s dividend”. This is the strategy in which a real criminal claims in court: “That recording of me taking a bribe? It's a deepfake, that's not me.” In a world where anything can be faked, authentic evidence loses its force because doubt seeps into every pore of the proceedings. The judiciary still lacks standardised protocols for the “AI authentication” of evidence, creating the risk that the innocent are convicted and the guilty acquitted because of “reasonable doubt” about a digital recording.

Automated cyber-attacks: crime that never sleeps

Classic hackers had to search for vulnerabilities manually. Today's AI bots scan thousands of servers per second looking for weaknesses. They learn from every failure. This is no longer a fight of person against person, but of our passive defence systems against autonomous malicious code.

As lawyers, we often meet clients whose data has been “locked” (ransomware), where AI helps criminals set the exact ransom amount by analysing the victim's financial position from publicly available data. It is a precisely targeted digital extortion attack.

A legal vacuum: are we a decade behind?

Although the AI Act is being adopted in Europe, in practice we still rely on laws written at a time when the “height of technology” was a portable hard drive. There are three key points where the law must evolve:

Liability of developers and platforms: who is responsible when AI deceives a person? The programmer who wrote the code, the company that deployed it, or the person who used it?

Cloud jurisdiction: high-tech crime happens “everywhere and nowhere”. We need international digital tribunals that can act faster than classic requests for international legal assistance, which travel for months.

The right to digital integrity: the Constitution must begin to protect our “digital footprint” as strictly as our physical home.

Prevention: the lawyer as a digital adviser

In the absence of a perfect law, your only defence is digital hygiene. As your legal representative, I advise:

• Multi-factor authentication (MFA): not an option, but an obligation.
• Verification protocols: never make financial transactions on the basis of a video or audio call without a second, independent channel of confirmation.
• Encrypted communication: protect your data before it becomes food for someone's AI model.

Conclusion: justice must grow digital muscles

Technology develops exponentially, while law develops linearly. That gap is the space in which digital predators thrive. Our role as lawyers is to be guards at that frontier, insisting on new laws that recognise AI not only as a tool but as a potential weapon.

The future lies not in banning technology, but in taming it through firm legal frameworks and technological literacy. Until then, we remain vigilant in a world where the image of you on a screen may not be you at all.

This text expresses the author's professional opinion and does not constitute legal advice for a specific case.