Deepfakes Hit 8 Million. Courts Still Can't Trust the Evidence. | Podcast

Full Episode Transcript

In 2023, roughly half a million deepfakes circulated online. By 2025, that number hit eight million. And according to U.N. Women, ninety-eight percent of them are non-consensual pornographic images targeting women.

Trusted by Investigators Worldwide

Run Forensic-Grade Comparisons in Seconds

Detailed facial comparison reports. Results in seconds.

Get Started

7-day refund guarantee**

That growth didn't happen in a vacuum

That growth didn't happen in a vacuum. Voice cloning now needs just a few seconds of audio to produce a copy so convincing it carries natural rhythm and emotional tone. Real-time face synthesis runs on consumer hardware. Yet according to U.N. researchers, fewer than half the countries on earth even have laws addressing online abuse — let alone laws written for A.I.-generated fakes. So who's supposed to hold anyone accountable when the legal system can't keep up with the toolbox?

Start with one investigator building a case. She's got a detection tool that flags a manipulated video. The algorithm is confident. But when she walks that evidence into a courtroom, the judge asks a simple question — can you explain how this tool reached its conclusion? Many of these detection methods are proprietary. The vendor won't disclose the inner workings. And under the Daubert standard — the legal test U.S. courts use to decide whether expert methodology is admissible — the judge needs to know if the method is testable, peer-reviewed, and carries a known error rate. A black-box algorithm struggles to clear that bar.

Meanwhile, money is flooding into the detection industry. According to Deloitte, that market is growing about forty-two percent a year, on pace to reach nearly sixteen billion dollars by 2026. But investment in detection doesn't solve the courtroom credibility problem. No standardized training or certification platform exists for analysts who compare faces in forensic settings. Every expert witness is essentially freelancing their own methodology.

The Bottom Line

And the human cost compounds that gap. When a survivor of deepfake abuse comes forward, the realism of the fabricated images makes them extraordinarily hard to disprove. According to U.N. Women, gender stereotypes can undermine a woman's credibility before she even presents evidence. She ends up defending herself against fabricated material — a second victimization layered on top of the first. The legal system asks her to prove a negative while the tools that could help her can't survive cross-examination.

The real divide isn't between real and fake anymore. It's between proof and credibility. Investigators can detect a deepfake — but detection is an opinion until a court calls it a fact.

So the picture looks like this. Deepfakes multiplied sixteen-fold in two years. Detection tools are a booming industry, but courtrooms still don't have a shared standard for admitting what those tools find. The investigators who'll matter most in the next few years won't just have the best algorithms. They'll have explainable methods — documented steps, confidence scores, audit trails — that a judge can actually evaluate. Watch for whether courts start adopting uniform admissibility frameworks for synthetic media. That's the bottleneck everything else is waiting on. The full story's in the description if you want the deep dive.

Deepfakes Hit 8 Million. Courts Still Can't Trust the Evidence. | Podcast