A Washington state superior court judge recently rejected the admission of video exhibits “enhanced by artificial intelligence” as evidence in a triple murder case.
To begin with, the state of Washington charged Defendant Joshua Puloka with three counts of murder stemming from a 2021 shooting. The shooting was captured on a bystander’s smartphone and the unaltered 10-second-long source video of the shooting had been entered into evidence.
The defense, however, also sought to admit an AI-enhanced version of the video which is when the Court heard testimony from defense witness Brian Racherbaeumer and State’s witness Grant Fredericks regarding proposed defense video exhibits enhanced by artificial intelligence.
Video Expert Witnesses
Brian Racherbaeumer is a self-identified videographer and filmmaker who started working with video in 1993. He was very candid and open about the fact that he is not – and has not claimed to be – a forensic video technician and has not been forensically trained.
Grant Fredericks is a contract instructor of video sciences at the FBI National Academy and is one of the most experienced video experts in North America. Moreover, Grant is a Certified Forensic Video Analyst, who has testified as an expert witness over 150 times in courts at all levels. He is recognized as a leading instructor in the science of Photographic/Video Comparison, Reverse Projection and Vehicle Speed Analysis.
Discussion by the Court
Racherbaeumer contended that the source video was low resolution, had substantial motion blur, and contained fuzzy images with “blocky” edge patterns. To fix these problems, Racherbaeumer stated that he had added clarity to the source video though the use of an AI-video editing tool in the Topaz Labs AI program before processing the video using an Adobe program. Racherbaeumer stated that the Topaz Labs AI program used technology that “intelligently scaled up the video to increase resolution,” as well as added sharpness, definition, and smoother edges to objects in the video.
However, the state challenged the proffered AI-enhanced video, asserting that it failed to meet the admissibility standard set forth in Frye v. United States – a standard requiring that evidence using novel scientific theories or principles must have achieved general acceptance in the relevant scientific community. According to Fredericks, the accepted forensic analysis of the video was impossible because of the AI tools used by the defense. Fredericks provided a litany of issues with the AI-enhanced video:
- the video added 16 times the number of pixels as existed in the original video, using an algorithm and enhancement method unknown and unreviewed by any forensic video expert,
- AI process removed information that was in the original images and it added information that was not in the original images,
- The proffered AI-enhanced video removed artifacts on individual images, and altered shapes and colors in the video. It removed the opportunity to forensically analyze which frames in the video utilized reference, predictive, and bi-directional images.
Also, Fredericks testified that the Scientific Working Group on Digital Evidence, whose members represented state, local, and federal law enforcement agencies engaged in forensic video examinations, had issued warnings regarding the use of AI-enhancement tools in the courtroom instead of approving the use of such AI-enhancement tools.
Held
The Court held that the Topaz Video AI enhancement tools, which use machine-learning algorithms, have not been peer-reviewed by the forensic video analysis community, are not reproducible by that community, and are not accepted generally in that community. Racherbaeumer himself was unable to say whether the Topaz Video AI he used, which had been commercially available for about three years, is currently utilized by the forensic video analysis community.
Moreover, the Court found that AI-enhanced video failed to satisfy Washington Rule of Evidence (ER) 702, under which evidence may only be admissible if it is reliable and will assist the trier of fact, as well as ER 403, which states that evidence is not admissible if its probative value substantially outweighs the danger of unfair prejudice.
Key Takeaway:
In conclusion, the Court held that the Defendant had the burden to show that the method was accepted in the relevant community because using AI tools to enhance video introduced in a criminal trial was a novel technique. The Court, as a result, held that the defense had failed to meet its burden after it found that the relevant scientific community was the “forensic video analysis community.”
Case Details:
| Case Caption: | State of Washington v. Puloka |
| Docket Number: | 21-1-04851-2 |
| Court Name: | Superior Court of Wasington for King County |
| Order Date: | March 29, 2024 |
Leave a Reply