In what appears to be a matter of first impression, a Washington State Superior Court judge recently refused permission to exhibit video “enhanced by artificial intelligence” for use in jury trials. in Washington State vs. Proca, Washington state has indicted Joshua Proka on three counts of murder stemming from a 2021 shooting. The shooting was captured on a bystander's smartphone, and a 10-second unaltered source video of the shooting was recorded as evidence.
However, the defense also asked to submit an AI-enhanced version of the video. The defense's experts argued that the source video was low resolution, had significant motion blur, and contained fuzzy images with “blocky” edge patterns. To address these issues, the defense's experts said they used the Topaz Labs AI program's AI video editing tools to enhance the clarity of the source video, and then further processed the video using an Adobe program. The defense's experts said the Topaz Labs AI program used techniques to “intelligently enlarge and increase the resolution” of the video, adding clarity, definition, and smooth edges to objects in the video.
The state objected to the proposed AI-enhanced video, arguing that it did not meet acceptable standards set by law. Frye v. United States – A standard that requires that evidence using new scientific theories or principles must be generally accepted in the relevant scientific community. According to the state's expert (a certified forensic video analyst), the AI tools used by the defense made forensic analysis of the accepted video impossible. The state's expert pointed out a series of problems with the AI-enhanced video.
- The video was enriched with 16 times the number of pixels than in the original video using algorithms and enhancement methods that are unknown and have not been verified by forensic video experts.
- I added some information that wasn't in the original file.
- Remove artifacts from individual images,
- The shape and color have been changed in the video.
The state's experts also testified that the Scientific Working Group on Digital Evidence, whose members represent state, local and federal law enforcement agencies involved in forensic video examination, had issued warnings about the use of AI-enhanced tools in courtrooms.
After hearing oral argument from both sides, the court rejected the defendants' proposal for AI-enhanced video, finding that the proposed evidence did not meet the requirements. fly standard. The court first held that because the use of AI tools to enhance videos submitted in criminal trials is a novel technology, the defendant has a responsibility to demonstrate that the method is accepted by the relevant community. I pointed it out. Determining that the relevant scientific community was the “forensic video analysis community,” the court found that the defense had not met its responsibility.
Specifically, the court held that the Topaz Video AI-enhanced tool, which uses machine learning algorithms, has not been peer-reviewed by the forensic video analysis community, is not reproducible by that community, and is not generally accepted by that community. Did. The court further noted that the defense has not presented any state or federal appellate decisions that have reviewed or approved AI-enhanced video in criminal or civil trials. The court also noted that the defense had not provided any articles, publications, or secondary legal sources authorizing the introduction of AI-enhanced video evidence in criminal or civil trials. The court also acknowledged that the defendant's experts themselves did not know what kind of videos the AI models were trained on, nor did they know whether such models employed generative AI in their algorithms. It also noted that it agreed the algorithm was “opaque and proprietary.”
Additionally, the court determined that the AI-enhanced video did not satisfy Washington Rule of Evidence (ER) 702. This rule states that evidence is only admissible if it is reliable and helpful to the trier of fact, and ER 403 provides that evidence is admissible only if it is reliable and helpful to the trier of fact, and ER 403 provides that evidence may only be admitted if its probative value significantly outweighs the risk of unfair prejudice. It is considered inadmissible as evidence. The court found that the AI-enhanced video did not honestly show what actually happened, but instead represented what the AI model thought it should show in an opaque manner, and that it We have determined that there is a significant risk that a lengthy trial will be held in a trial that involves a process that does not exist.
The court's decision Proca This is another example of law enforcement skepticism about allowing AI-enhanced or generated evidence to help determine the outcome of a case. procaCourts have noted on several occasions a lack of transparency regarding the inputs of AI models, and unless courts are provided with more information about such inputs, they are likely to continue their tendency of omission.
