By When Paris Hilton addressed the throng of press on Capitol Hill earlier this year, it wasn’t to punt a new product or reality show. Her reason was about an issue that might soon affect the global population.
She used her position of influence and power to advocate for the DEFIANCE Act. The bipartisan U.S. legislation targets non-consensual sexually explicit deepfakes. Passed by the Senate in January, the bill empowers victims to sue anyone who creates, distributes, or shares non-consensual sexually explicit deepfakes.
Using her own experience as a catalyst for the passing, Hilton recounted the trauma that ensued following the release of her leaked sex tape in 2004. She was only 19 at the time. But her nightmare was far from over. AI deepfakes infiltrated the internet; 100,000 explicit images of her circulated online without her consent.
“They sold my pain for clicks, and then they told me to be quiet, to move on. To even be grateful for the attention.” – Paris Hilton while addressing Capitol Hill.
The scary part is that fame doesn’t make you a target. According to the Henry L. Stimson Center, about 98% of all deepfake content is non-consensual and pornographic. And 99% of those depicted are women.
That’s one part of the puzzle. The other is proving that those hyper-realistic images are not real. Below, we’ll discuss the legal risks of deepfake evidence in civil litigation.
A Solid Base
Attorneys with extensive knowledge of deepfakes and AI are the best-positioned to take on these cases. So are newly graduated and licensed attorneys.
Previously, law school concentrated on traditional evidence: eyewitnesses, forensics, etc. Deepfakes disrupt the status quo. Intellectual property and privacy law are no longer black-and-white. That gray area is called the epistemological crisis in courtrooms.
When studying towards a JD law degree, students study deepfakes in several core subjects:
- Evidence and Procedure
- Intellectual Property and Rights of Publicity
- Criminal and Civil Torts
- Legislation and Compliance
Many choose to study an online Juris Doctor law degree part-time with minimal disruption to their professional careers. Cleveland State University advises enrolling in an ABA-accredited online JD program.
The Admissibility Trap
AI forensics within the legal framework is a moot point. While existing computer forensics principles are being adapted to fit the emerging field, the Legal Intelligencer explains that a universal legal standard has yet to be established.
Because of this oversight, AI forensic methods could face challenges in court due to the absence of a standardized benchmark.
Technically speaking, most deepfake detection tools indicate an 86% probability that the video is a deepfake. However, these scores are frequently presented without context. In traditional digital forensics, “likely” conclusions are supported by clear evidence.
Without proper context, it’s harder to explain the findings and determine how much weight AI-generated forensic conclusions should carry in court.
Seeing is Believing
There are psychological aspects at stake here as well. Psychology Today reports that we are wired to believe what we see. Our minds are bad at detecting fake evidence, says practicing attorney Ruth S. Johnson.
Researchers have coined a term for it: veridicality heuristic. It’s why video evidence has always been persuasive. Like the Legal Intelligencer’s argument, Johnson agrees that technologies that detect AI-generated content are either unreliable or biased.
What defense attorneys are doing is invoking the “deepfake defense.” They contend that authentic recordings could be AI-generated fabrications. The result is a courtroom double bind. Jurors may doubt genuine evidence while remaining vulnerable to convincing fakes.
FAQs
1. What is the DEFIANCE Act?
The DEFIANCE Act is bipartisan U.S. legislation aimed at combating non-consensual sexually explicit deepfakes.
2. Why are deepfakes becoming a challenge in civil litigation?
Deepfakes blur the line between authentic and fabricated evidence. Courts must determine whether images, audio, or videos are genuine.
3. Can current AI tools reliably detect deepfakes?
Not entirely. Many deepfake detection tools assign a probability score, but these results are sometimes presented without sufficient context or supporting evidence.
4. Why is legal education increasingly focusing on deepfakes?
Future attorneys need to understand its impact on evidence, intellectual property, privacy rights, civil liability, and criminal law.
Deepfake Litigation: Key Statistics
| Statistic | Figure | Why It Matters |
| Deepfake content that is non-consensual and pornographic | 98% | The overwhelming majority of deepfakes involve sexual abuse and exploitation. |
| Victims depicted in explicit deepfakes who are women | 99% | Women remain the primary targets of non-consensual deepfake pornography. |
| Probability score commonly produced by deepfake detection tools | 86% | Detection tools estimate the likelihood that content is AI-generated rather than providing definitive proof. |
| Age of Paris Hilton when her sex tape was leaked | 19 years old | Her experience has made her a leading advocate for stronger protections against digital abuse. |
Closing Argument
The law is bound to change as technologies evolve. Yet it isn’t changing fast enough. Presenting deepfake evidence in civil litigation is proof of this. Hopefully, the DEFIANCE Act will provide some sense of protection against digital abuse.
The next generation of attorneys who’ve graduated from a Juris Doctor program will be well-equipped to tackle these cases. And AI detection tools won’t leave any room for doubt. In the meantime, victims have to trust that the law will run its course and in their favor.
