We use cookies on this website. By continuing to use this site without changing your cookie settings, you agree that you are happy to accept our privacy policy and for us to access our cookies on your device.
Transcript/ScriptUSAGM SHARE
((PLAYBOOK SLUG: Elections AI Deepfakes
HEADLINE: AI Deepfakes Pose Threat to Elections Worldwide
TEASER: Manipulated videos and audio could sway votes in some 50 elections around the world
PUBLISHED:
BYLINE: Valdya Baraputri
CONTRIBUTOR: Naras Prameswari
DATELINE: Washington
VIDEOGRAPHER:
VIDEO EDITOR:
ASSIGNING EDITOR: Ade Astuti
SCRIPT EDITORS: KEnochs; Reifenrath
VIDEO SOURCES: VOA original, AFP, AP, REUTERS, Courtesy video, Zoom with License
PLATFORMS: WEB __ TV _x_ RADIO __
TRT: 3:11
APPROVED BY: KE
TYPE: TV
EDITOR NOTES:)
((INTRO))
[[An estimated four billion people worldwide have or are scheduled to vote in national elections through 2024. While the candidates and issues differ in each country, one big concern is uniting vote watchers. VOA’s Valdya Baraputri has the story.]
((NARRATION))
In the weeks leading up to Indonesia's February 14th elections, voters have been inundated with manipulated video and audio — or deepfakes — of politicians … like this video of the late Indonesian President Suharto spread via social media. ((Mandatory courtesy: TikTok / @abialasyi))
Fact-checkers like Santi Indra Astuti of the Indonesian Anti-Defamation Community — Mafindo for short — say the use of generative artificial intelligence is what sets disinformation campaigns apart in 2024.
((RADIO TRACK: She says that during the 2014 election, most hoaxes were used to discredit candidates. In 2019, they began to be used to discredit election organizers. Now, in 2024, hoaxers are attempting to discredit the entire democratic process))
((Santi Indra Astuti, Mafindo Digital Literacy Activist (female, in Indonesian))) ((Zoom))
"During the 2014 election, most of the hoaxes were used to discredit the opposing candidates. In 2019, hoaxes began to discredit election organizers. While in 2024, we've seen an even more bizarre escalation, to discredit the whole democratic process."
((NARRATOR)) ((Mandatory courtesy: TurnBackHoax.id))
On its website Turn Back Hoax, Mafindo revealed a deepfake phone conversation in which Indonesian Democratic Party of Struggle boss Megawati Sukarnoputri harshly reprimands presidential candidate Ganjar Pranowo.
((Mandatory courtesy: TurnBackHoax.id))
Other videos show presidential hopefuls Anies Baswedan and Prabowo Subianto fluently speaking Arabic, which none of them do in reality, to appeal to the nation’s Muslim voters.
((RADIO TRACK: Santi of Mafindo says it appears hoaxes are being used not only to discredit opponents but also to glorify candidates.))
((Santi Indra Astuti, Mafindo Digital Literacy Activist (female, in Indonesian)))
((Zoom))
"Hoaxes are no longer just to discredit opponents, but some seem intended to glorify their chosen candidate."
((NARRATION))
The U.S. has just started primary elections, ahead of general elections in November, and already AI is posing a challenge. In one doctored robocall, President Joe Biden appears to encourage Democrats to stay home from the polls.
[[for Radio: Lindsay Gorman is an expert on emerging technologies at the German Marshall Fund's Alliance for Securing Democracy]]
((Lindsay Gorman, German Marshall Fund (female, in English)))
((Zoom))
“AI-generated media, from robocalls to deepfake audio to campaign ads, are being featured in almost every election that we're seeing both in the end of 2023 and now into 2024, so I think this is becoming the new norm for our political and electoral landscape”
((NARRATOR))
Most voters who spoke to VOA say they are not confident that they can tell deepfakes and real content apart.
((Shae Hooper, US Voter))
“So we’re just hoping that news does their job. And telling us what’s real and fake."
((Felipe Villegas, US Voter)) “And there’s so many random people doing those things that it’s going to be near impossible to stop”.
((NARRATOR))
Safeguards are being put in place at the federal level. President Biden signed an executive order in October urging government agencies to issue guidelines designed to protect consumers from A-I hoaxes. But such measures have not been that effective, Gorman says.
((Lindsay Gorman, German Marshall Fund (female, in English)))
((Zoom))
“Right now there is no requirement that actually we have to label that piece of content as AI generated. Even if it is a piece of content used for political campaign."
((NARRATOR))
AI company DeepMedia, contracted by the US Department of Defense to detect deepfakes, estimates a half a million deepfake audio or video clips were spread globally through 2023. Analysts fear that will only go higher during this politically charged year.
((Valdya Baraputri, VOA News, Washington.))
NewsML Media TopicsArts, Culture, Entertainment and Media
NetworkVOA
Embargo DateFebruary 8, 2024 12:26 EST
BylineValdya Baraputri
Brand / Language ServiceVoice of America - English, US Agency for Global Media