How Generative AI and Deepfakes Are Challenging the Criminal Justice System

The Emerging Threat to Evidence, Trust, and Digital Reality

April Liuzzi - March 18, 2026

Reality can no longer be trusted at face value.

Abstract

Generative artificial intelligence (AI) and deepfake technologies are rapidly transforming the nature of crime and posing new challenges for the criminal justice system. This paper synthesizes recent research (2020–2026) to examine how AI-enabled tools are expanding the scale, accessibility, and sophistication of criminal activity. Findings suggest that generative AI lowers barriers to entry for cybercrime, accelerates fraud and impersonation schemes, and undermines the reliability of digital evidence. Deepfakes, in particular, introduce a growing risk to legal systems by blurring the line between authentic and manipulated media. The analysis further highlights how criminological theories—specifically Routine Activity Theory and Rational Choice Perspective—help explain the rise of AI-driven crime. The paper concludes that while policy responses are emerging, technological advancement continues to outpace regulation, creating a widening gap between innovation and institutional readiness.

Introduction

What happens when seeing is no longer believing?

Generative AI has made it possible to create highly realistic text, images, audio, and video—often indistinguishable from reality. While these tools have valuable applications, they also introduce a new category of risk: the ability to fabricate convincing digital content at scale.

One of the most concerning outcomes is the rise of deepfakes—AI-generated or manipulated media that can imitate real people, events, or voices with alarming accuracy. What once required advanced technical expertise can now be done with minimal effort.

This shift is not just technological—it is systemic.

Criminal activity is becoming:

• Easier to execute

• Harder to detect

• More difficult to prosecute

This paper explores how generative AI is reshaping crime and examines what this means for the future of the criminal justice system.

The Growing Scope of AI-Driven Crime

The scale of AI-enabled crime is expanding rapidly.

Recent data shows that deepfake-related scams increased by over 1,300% in a single year, with a significant portion of high-value fraud involving synthetic media. This growth is driven by one key factor: Accessibility.

AI tools have lowered the barrier to entry for cybercrime. Tasks that once required specialized knowledge—such as creating fake identities or manipulating media—can now be performed with basic tools and limited experience.

Why This Is Happening

Several forces are driving this trend:

1. Easy Access to AI Tools

Even inexperienced individuals can now generate convincing scams, impersonations, or fraudulent content using widely available AI systems.

2. Abundance of Personal Data

Social media provides a constant stream of photos, videos, and voice recordings. This data can be used to create highly personalized deepfakes.

3. Increased Attack Efficiency

AI enables:

• Faster execution of scams

• More convincing deception

• Larger-scale targeting

4. Expanding Organizational Risk

Businesses are increasingly vulnerable.AI-related breaches are rising, with millions of records exposed in major incidents and significant financial losses.

Deepfakes and the Collapse of Trust in Evidence

One of the most serious implications of generative AI is its impact on evidence.

Digital media—videos, audio recordings, images—has long been a cornerstone of criminal investigations. But deepfakes challenge a fundamental assumption: That recorded evidence reflects reality.

The “Liar’s Dividend” Problem

As deepfake awareness grows, a new issue emerges: People can deny real evidence by claiming it is fake.

This phenomenon—often called the “liar’s dividend”—creates a dangerous situation where:

• Fake evidence can appear real

• Real evidence can be dismissed as fake

The result is a loss of trust, not just in media, but in the justice system itself.

Why Criminals Are Using AI

Two major criminological frameworks help explain this shift:

Routine Activity Theory (RAT)

Crime occurs when three elements align:

• A motivated offender

• A suitable target

• Lack of effective guardianship

AI strengthens all three:

• Offenders gain powerful tools

• Targets are easily identified online

• Guardianship (cybersecurity, enforcement) is often weak

In digital environments, these conditions are not rare—they are constant.

Rational Choice Perspective (RCP)

Criminals make decisions based on risk vs. reward.

AI changes that equation:

• Lower cost and effort

• Higher potential rewards

• Reduced likelihood of detection

Deepfakes, for example, allow criminals to impersonate trusted individuals, making scams more effective and harder to identify.

Figure 1. Growth of AI-Enabled Crime Risk (Conceptual Model)

Indexed growth of AI-enabled criminal activity based on reported trends

Policy Responses and Their Limits

Governments are beginning to respond—but slowly.

Key Policy Areas

• Criminalization of non-consensual deepfakes

• Regulation of political AI-generated content

• Platform accountability laws

Some legislation requires:

• Removal of harmful AI-generated content

• Disclosure of manipulated media

The Core Problem

Technology is evolving faster than regulation.

Legal systems are reactive by nature, while AI development is rapid and continuous. This creates a persistent gap between:

• What technology can do

• What laws can control

Technological Countermeasures

There is growing investment in:

• Deepfake detection tools

• AI-based verification systems

• Digital authentication frameworks

However:

• Detection is not fully reliable

• Tools struggle to keep pace with new AI models

What This Means for the Future

This is bigger than cybercrime—it’s about trust in reality itself.

1. Evidence Will Require Verification

Courts may need new standards for authenticating digital media.

2. Crime Will Become More Scalable

AI allows fewer individuals to commit more crimes at a larger scale.

3. Public Trust May Erode

If people cannot distinguish real from fake, confidence in institutions may decline.

4. New Skills Will Be Required

Law enforcement and legal professionals will need:

• Technical expertise

• Advanced forensic tools

• Cross-sector collaboration

Discussion

The rise of generative AI represents a structural shift in how crime is committed and understood. It does not simply enhance existing threats—it fundamentally changes them.

The criminal justice system faces a dual challenge:

• Adapting to new forms of evidence manipulation

• Maintaining legitimacy in an environment of uncertainty

Without significant investment in detection, regulation, and education, these challenges may intensify.

Conclusion

Generative AI and deepfake technology are redefining the boundaries of crime, evidence, and trust.

While these tools offer significant benefits, their misuse presents serious risks to the integrity of the criminal justice system. The increasing realism of synthetic media challenges long-standing assumptions about truth and authenticity.

The central issue is not just technological—it is institutional. The future of justice may depend on our ability to determine what is real.

References

Department of Homeland Security. (2024). Impact of artificial intelligence on criminal and illicit activities: 2024 public-private analytic exchange program.

Dzuba, C. (2025). Artificial intelligence and social engineering attacks: Applying routine activity theory to emerging cyber threats. Issues in Information Systems, 26(2), 452–465. https://doi.org/10.48009/2iis134

Fortinet. (n.d.). How deepfake AI is transforming cybersecurity. Retrieved March 14, 2026, from https://www.fortinet.com/resources/cyberglossary/deepfake-ai

Holt, T.J., & Bossler, A.M. (2009). Examining the applicability of lifestyle-routine activities theory for cybercrime victimization. Deviant Behavior, 30(1), 1-25. https://doi.org/10.1080/01639620701876577

Olmez, S., Birks, D., Heppenstall, A., & Ge, J. (2024). Learning the rational choice perspective: A reinforcement learning approach to simulating offender behaviours in criminological agent-based models. Computers, Environment and Urban Systems, 112, Article 102141. https://doi.org/10.1016/j.compenvurbsys.2024.102141

Office of Communications. (2024). Traditional AI vs. generative AI: What’s the difference? College of Education, University of Illinois. https://education.illinois.edu/about/news-events/news/article/2024/11/11/what-is-generative-ai-vs-ai

Rieper, M. (2026). How AI-generated content laws are changing across the country. MultiState.

Roscoe, J. (2025). Deepfake scams are distorting reality itself. Wired.

Rumage, J. (2026). These 12 major tech laws take effect in 2026. Here’s what you need to know. Built In.

Tulga, A.Y. (2026). AI-driven fraud as an emerging cyber risk: Evidence from a global incident-based analysis. EDPACS. https://doi.org/10.1080/07366981.2026.2631066

Sandoval, M. P., de Almeida Vau, M., Solaas, J., & Rodrigues, L. (2024). Threat of deepfakes to the criminal justice system: A systematic review. Crime Science, 13(1). https://doi.org/10.1186/s40163-024-00239-1

Schiff, K. J., Schiff, D. & Bueno, N. S. (2024). The liar’s dividend: Can politicians claim misinformation to evade accountability? American Political Science Review. https://doi.org/10.1017/S0003055423001454

Steinberg, S. (2020). How deepfakes harm the criminal justice system. South Jersey Legal Blog. https://www.zswlaw.com/blog/2020/07/how-deepfakes-harm-the-criminal-justice-system#:~:text=Deepfakes%20affect%20the%20entire%20criminal,lead%20to%20it%20being%20excluded.

Williams, M. (2016). Guardians upon high: an application of routine activities theory to online identity theft in Europe at the country and individual level. The British Journal of Criminology, (56)1, 21-48. https://doi.org/10.1093/bjc/azv011