FEED Winter 2024 Web

that misinformation will affect what they see. A key point to reiterate is that this form of misinformation isn’t new. I’m more concerned about claims of things being fake that use AI for plausible deniability.” Unfortunately, Leibowicz’s concerns surrounding this reverse-psychology approach were far from unfounded. During Vice President Kamala Harris’ rally in Detroit back in August, Trump falsely claimed that images of swarming crowds were AI-generated, thus fanning the disinformation flames among Republican supporters. “I think our solution to the misinformation problem needs to be privy to the fact that it’s possible to manipulate public opinion without using technology. “I come to this from a background of technology within news,” says MacCormack. “Working with CBC Radio Canada, we started worrying about AI around 2017/18. We quickly concluded that stand-alone detection wasn’t a viable solution for many reasons – mainly not having enough longevity and that the accuracy isn’t going to scale. “A key approach is to follow a file’s history and prove what is true. They have been doing this in banking for years, so we thought we’d do a lift and shift. Eventually we created something called the Coalition for Content Provenance and Authenticity, which is now 120 companies strong.” A collaborative, inter-sector approach to verification is crucial. Additionally, many media organisations have been known to develop internal initiatives. BBC Verify, for example, is dedicated to examining the facts and figures behind a story to

determine whether or not it’s true. It works across the entirety of BBC News and now has a consistent presence throughout its election coverage. One of the panellists, Mounir Ibrahim, shares his company Truepic’s impact within the validation arena. “We started thinking about this back in 2015/16, pre-AI,” he says. “I was formerly a US diplomat and saw images coming from war zones daily. We often couldn’t determine if they were accurate, so the question was always: if we can’t prove that it’s fake, how can we prove that it’s real? “That’s how Truepic came about. We felt that if we could figure out a way to cryptographically hash and secure metadata and pixels from a point of capture, we could then determine an interoperable standard that can allow verified information to travel safely through the internet.” MacCormack emphasises the importance of organisations like Truepic, stating that “as of last year, 150 billion synthetic images have been created, which is the same number as that of photographs taken since the invention of the camera.” EXTENDED ELECTIONS Thinking about election coverage, the mind conjures up images of glossy studios featuring presenters interacting with screens, wandering across AR maps and swiping through virtual data. We’re, of course, discussing extended reality – a dominant force in most kinds of news coverage at present. The global XR market reached almost $30

MAKING HEADLINES The Washington Post is just one of the news outlets that Brainstorm services with graphics and on-screen data analysis

I’M MORE CONCERNED ABOUT CLAIMS OF THINGS BEING FAKE THAT USE AI FOR PLAUSIBLE DENIABILITY

feedmagazine.tv

Powered by