�山����

By Joe Arney
Photos by Kimberly Coffin (CritMedia, StratComm’18) and Nandi Pointer (PhD candidate, Media Studies)

Ask any of her students how they prefer to get their news, or search for recommendations, or learn about their favorite TV shows, and Sandra Ristovska will tell you that they go on TikTok.

Yet their educations—from the time they first set foot in a grammar school classroom—have focused on textual literacy, with almost nothing devoted to how video and photos are analyzed.

“We just assume that everybody intuitively knows how to understand images, because we don’t have to teach you an alphabet, or grammar,” said Ristovska, associate professor of media studies at the College of Media, Communication and Information. “But we know from research that people can watch the same image and arrive at a vastly different understanding about what that image says or does.”

More: Bringing student activism to TikTok videos

That’s fun when we’re overanalyzing a plot twist in Severance. But Ristovska’s work centers around what happens when videos make their way into a courtroom, where interpretations can influence a person’s guilt or innocence.

According to one estimate, video appears in about 80 percent of criminal cases, but no guidelines exist for how video can be presented as evidence in court. And that’s also the case for deepfake videos or media created by generative artificial intelligence.

“Anybody who’s seen a legal document knows they’re standardized—if it doesn’t look a certain way, it’s not going to be admissible in court,” Ristovska said. “But when it comes to video, different courts have different guidelines and understandings about what’s admissible.”

Ristovska has been an important contributor to scholarship in media and the law. At a daylong event in April, she helped steer the conversation around these topics while formally presenting the Visual Evidence Lab, a new lab at CMCI that will advance her work in this area.

The workshop, Justice by Video, brought together judges, attorneys, journalists, and scholars from the humanities, social sciences, law and STEM to develop new avenues for research and potential policy proposals around how to ensure justice is best served.��

Ristovska’s personal history plays a role in all this, too. Growing up in what is now Macedonia during the Yugoslav Wars, she still recalls how footage from the fighting upset her parents—even if she was too young to understand the news bulletins interrupting her evening cartoons. As part of her graduate school work, she went on to study how footage from civilians and activists made its way to the United Nations’ criminal tribunal, in The Hague.

“I realized the law was an important place to be asking questions about video evidence,” she said. “Some of the citizen footage in the tribunal wasn’t verified through the person who shot it, which had never been the case before. And this footage was both establishing the truth in court while constructing a historical memory about the wars.”

Cross-disciplinary expertise

Sandra Braman, a professor of media and information at Michigan State University, said she was excited to participate in this event because of the range of expertise involved, including practicing judges as well as legal scholars and researchers from across the social sciences.

Braman has twice served as a visiting professor at CMCI, and is considered among the leading scholars in digital technologies and their policy implications. She was impressed with the agenda, which included small group discussions intended to stimulate cross-disciplinary discussion and a detailed reading list to review beforehand.

“Usually, when you go to the first conference of its kind, it’s just a chance to gather and talk generally about the topic,” Braman said. “Sandra has put together a very structured set of tasks that are actually very hard questions to guide us on visual evidence.”

Roderick Kennedy, who retired from the New Mexico Court of Appeals after serving as its chief judge, was part of an afternoon panel discussing the issues raised by and the role security footage plays in creating a narrative explaining what happened.

Kennedy and Ristovska met through his work with the American Bar Association. Ristovska presented a series of webinars on video evidence and deepfakes to members. They also collaborated when she was a guest editor of last winter that took a deep dive on these issues.��

Kennedy said video evidence presents similar challenges that he would see with eyewitness testimony throughout his career. Memory is unreliable, he said, as witnesses become suggestible when asked to remember details or are affected by the pressure to have a definitive answer for investigators.

“You have a single viewpoint, but it’s overlaid with other memories that can change things, and is subject to interpretation every time you recall it and restore it,” he said.

‘A vertical learning curve’

A video won’t change its memory under pressure, but how it’s captured and edited can influence the way a jury interprets what happened. And while footage from police body cams or the smartphones of bystanders may get the most attention, Kennedy said the issue crops up elsewhere—even police interrogations. He shared a case involving a pathologist whose findings in a homicide were influenced by hearing a woman confess to the crime on camera.

Her confession, however, was preceded by an exhausting, seven-hour police interrogation. And because we’ve been conditioned to believe videos show reality—without considering how they were framed, trimmed, slowed down or otherwise edited—they have significant potential to mislead jurors.

“That’s the power of video,” Kennedy said. If you only show a jury the last minute or so of that interrogation, “all you see is a mother saying she killed her baby.”

The workshop wasn’t just about editing techniques that may introduce doubt. Invited experts also discussed deepfakes, an emerging challenge for courts that must catch up to the technology. Kennedy said judges and lawyers “have almost a vertical learning curve” when it comes to the technology.

“You have to learn the language of the technology experts before you can accuse somebody of using a deepfake,” he said. “And the experts aren’t taught how to speak legal, or the legal rules for putting their expertise in evidence.”

One thread of Braman’s research on information policy is the history of facts themselves.

“Our social orientation around facts provides the context within which we think about evidence,” she said. “And though we are talking a lot today about A.I. and the problem of deepfakes, the question of the authenticity and validity of digital information in general actually first arose as soon as the internet became available to the general public. We need to solve this problem yesterday.”

Ristovska said she was pleased to see members of the public attend to watch Incident and start thinking about video as a communication tool that is overdue for guidance.

“We’re not going to solve all the challenges around how people see video—we can’t do that with any type of evidence,” she said. “But I hope we can develop research-based guidelines that promote consistency, fairness and equality in the use of video as evidence.”