A public-private partnership in Utah aims to help campaigns combat deepfakes by authenticating the content they publish online.
Authentication of content is an area where some cybersecurity experts believe there is a gap in the campaign industry and startups have sprung up to meet that need.
For instance, Garrett Kinsman recently launched an app called Click to help campaigns authenticate digital images or videos. In Utah, academics at Utah Valley University (UVU) are working in partnership with local startup SureMark Digital Identity Services to provide content authentication services to each party-nominee seeking one of Utah’s four congressional seats and the open Senate seat.
Brandon Amacher, director of the Emerging Tech Policy Lab at UVU’s Center for National Security Studies, credited Amelia Powers Gardner, a Utah County commissioner, with helping spearhead the partnership.
“We were all concerned about [deepfakes] separately in our own lanes, but just kind of came to the recognition that it required a more holistic approach to really try to get out in front of,” Amacher told C&E.
The team up of academics, private business and campaigns is meant to create a chain of custody for digital content.
“This is for political campaigns to be able to digitally sign and authenticate real content, which is coming from their campaigns,” Amacher said. “Detection is a separate issue. This is more a way for a campaign to say, ‘Hey, we legitimately produced this video and now you, as the end user, as the end voter, you can verify [that] we did produce this.”
While the authentication of content has no financial barrier for the campaigns, it does require voters to use a free browser plug-in from SureMark Digital Identity Services in order for them to get that confirmation.
The university will act as a third-party evaluator for the pilot program with future research planned from the data that’s gathered this cycle.
“We’re going to be talking to these campaigns — we’re going to be getting feedback, collecting data, analyzing this model,” Amacher said. “We’re not necessarily endorsing one product or one model to address this issue. We want to really get out there, try something new, see where it work, where it does not work [and] how we can stack other solutions on top of it.”
