As artificial intelligence (AI) reshapes the legal system, courts are taking a closer look at digital evidence. This is the first article in a series on how AI is affecting criminal justice in Florida and the new tech risks defense teams must handle.
As synthetic media improves, deepfakes make audio and video less trustworthy. They can make it look like someone said or did something that never happened. For people facing serious charges, knowing how to challenge deepfakes is key to protecting their rights.
The authentication requirement in Florida
The primary battleground for AI-generated evidence is the Florida Evidence Code, specifically regarding authentication. Evidence must be checked to confirm it is what the party presenting it says it is.
Florida courts have often relied on the silent witness theory to admit recordings. But deepfakes show the theory’s limits because fake media can look real. Because of this, defense lawyers are disputing digital evidence more often. They are also asking for proof of a clear digital chain of custody that shows the file was not changed or altered by AI.
Real-world implications
The technology is still fairly new, but Florida’s legal community is on alert after cases of AI-generated misinformation. One major example was in 2024, when AI-generated audio led to a Maryland school principal facing a wrongful accusation of hate speech. In 2026, a California judge also found a deepfake video in the submitted testimony of the plaintiff in the Mendones v. Cushman & Wakefield case.
In criminal court, a deepfake confession or a fake video of a crime might be used as evidence. If the defense does not challenge it early, a jury could convict based on a digital hallucination.
Role of machine learning forensics
To counter synthetic media, criminal defense attorneys increasingly work with digital forensics teams to question evidence that may have been altered. Machine-learning forensics can spot AI traces or small mismatches left by deepfake tools. They may analyze the following:
- Biometric deviations: Irregularities in blinking patterns, pulse detection in skin tones or unnatural shadows
- Metadata: Odd details in the file’s metainformation that suggest the use of generative software
- Spectral patterns: Noise or frequency shifts in audio that are may point to AI-generated speech
By bringing these technical experts into the courtroom, the defense can ask the court to exclude evidence if the prosecution cannot prove it is real.
Protecting you from fake evidence
As technology evolves, so must the strategies for defending the accused and challenging AI-generated evidence. A dedicated defense lawyer can help ensure that proof has a basis on reality, not an algorithm.

