In the US state of Washington, a judge has ruled against the use of artificial intelligence (AI) enhanced video evidence in a triple murder case citing concerns over the “novelty” and “opacity” of the technology.
On Monday, the judge expressed concerns about the novelty and opacity of AI technology, particularly in understanding decisions related to AI-enhanced video content.
This ruling is unprecedented in US criminal courts and arose from a case where the defense sought to introduce mobile phone footage, enhanced by machine-learning software, in a defense related to a shooting incident outside a bar near Seattle.
Prosecutors objected to the technology’s use, pointing to a “lack” of legal precedent, marking a notable legal development regarding the admissibility of AI-enhanced evidence.
The defense employed machine learning, a form of AI, to enhance mobile phone footage in support of the defendant’s self-defense argument.
However, Topaz Labs, the software developer utilized by the defense, warned against employing its technology for forensic or legal purposes, underscoring its unsuitability for such applications.
A forensic video analysis by Frederick Grant revealed inconsistencies between the original and AI-enhanced videos, with experts noting inaccuracies and misleading elements in the AI-enhanced version.
Grant emphasized that while the AI technology may produce visually appealing results, the enhanced clarity and resolution might not accurately represent the events captured in the original footage.
Despite claims from the defense about the fidelity of the enhanced video, experts expressed concerns about the lack of established methodologies and research supporting the reliability of AI-based video enhancement.
Former crime scene investigator and forensic video analyst George Reis emphasized the need for thorough research and standardized procedures before widespread adoption of AI technology in legal contexts.
While AI has been explored to some extent for investigative purposes, such as enhancing license plate images, Reis stressed the importance of ensuring the reliability and validity of AI-enhanced evidence through rigorous testing and established standards.
“At this particular point, it’s premature to determine the appropriate level of AI use for clarifying still photographs or videos in legal contexts,” Reis commented.
YOU MAY ALSO READ: Zimbabwe declares state of disaster as El Nino-induced drought threatens food security