Experts worry about how AI deep fakes may impact 2024 US elections

In a special report by Reuters, various industry experts were asked their opinions on the looming issue of AI deep fakes for the upcoming 2024 US elections.
Christopher McFadden
us-2024-elections-deepfakes.jpg
Will deep fakes mire the 2024 US elections?

mrolands/iStock 

As reported by Reuters, next year's elections in the United States are likely mired with ever-sophisticated deepfake videos and audio. The rise of artificial intelligence (AI) tools like Midjourney and FakeYou have made it nye-on child's play to generate realistic-looking images, videos, and audio files that are very difficult at times to distinguish from reality.

According to Reuters, there has been a significant increase in the creation of "synthetic" media of late using new "generative AI" tools like Midjourney. This technology has been available for some time, but recent advancements have made producing convincing deep fakes easier and more affordable. To better understand the potential problem, Reuters gathered input from approximately 24 experts in various fields, including AI, political activism, and online misinformation, to determine how serious the AI deep fakes could become.

"[It will] be very difficult for voters to distinguish the real from the fake. And you could just imagine how either Trump supporters or Biden supporters could use this technology to make the opponent look bad," Darrell West, a senior fellow at the Brookings Institution's Center for Technology Innovation, told Reuters. "There could be things that drop right before the election that nobody has a chance to take down," he added.

According to Aza Raskin, co-founder of the Center for Human Technology, the tech sector's AI arms race has led to the release of deep fake generation tools without adequate safeguards to prevent the spread of harmful misinformation. Although Facebook, Twitter, and YouTube have attempted to ban and eliminate deep fakes, their success in regulating this content is inconsistent.

According to DeepMedia, a company developing tools to detect synthetic media, the number of video deep fakes and voice deep fakes posted online has tripled and octupled, respectively, compared to the same period last year. By their estimation, reports Reuters, approximately 500,000 videos and voice deep fakes will circulate on social media platforms worldwide by 2023. Until recently, it used to cost $10,000 for voice cloning in terms of server and AI training expenses. However, startups now provide this service for only a few dollars.

No one is certain where the generative AI road leads or how to guard against its power for mass misinformation effectively, but, like any media the public consumes, it is yet to be seen how much it will influence people's decisions. Political propaganda has been around since time immemorial, but faked videos, images, and audio "evidence" may affect people's decisions far more than previously possible.

That being said, many experts in various fields tend to share their thoughts on social media, like Twitter or YouTube, so any fakes will likely quickly be spotted and called out. We will have to see what emerges on the net in the run-up to the elections.

Some might even be comedy gold!

Add Interesting Engineering to your Google News feed.
Add Interesting Engineering to your Google News feed.
message circleSHOW COMMENT (1)chevron
Job Board