Trump supporters forge AI deepfakes to woo Black voters • The Register

Admirers of Donald Trump are creating fake AI-generated images depicting the former US president interacting with Black people, in the hope the pictures are interpreted by voters as a sign of increasing support from a key voter group that would bolster his attempt to regain the White House.

Anyone can create fake images with commercial generative AI tools. One such effort showed Trump surrounded by young Black men and, after being shared on X by a parody account and falsely described as showing the likely Republican nominee as having stopped his motorcade to meet them, racked up over 1.3 million views according to BBC Panorama

Another image shared the same account shows Trump standing at what appears to be a protest with his fist up and is captioned: “No one has done more for the black community than Donald Trump.”

These fake images look realistic. Their proliferation can be taken to suggest that Trump is campaigning in, and gaining popularity among, the African American community – a key demographic in any US election.

Mark Kaye, a popular conservative radio talk show host with over a million followers, used AI to make an image of Trump with Black people and shared it on Facebook. He admitted the picture was fake. “I’m not out there taking pictures of what’s really happening. I’m a storyteller,” he argued.

He didn’t believe that he was doing anything wrong by generating and spreading false images. “I’m not claiming it is accurate. I’m not saying, ‘Hey, look, Donald Trump was at this party with all of these African American voters. Look how much they love him!'”

“If anybody’s voting one way or another because of one photo they see on a Facebook page, that’s a problem with that person, not with the post itself,” he concluded.

The US government and social media companies are stepping up efforts to monitor and tackle political deepfakes.

“We do not have any specific or credible threat to today’s election operations” a senior official from the US government’s Cybersecurity and Infrastructure Security Agency said in a briefing on Tuesday.

“We’ve put a lot of efforts in that we’re focusing on the increased risk posed by generative AI capabilities in this space. That’s a threat vector that we focused on over the past several months to try to encourage election officials across the country to understand the threat and the steps they can take to mitigate against it.”

Meanwhile, corporations like OpenAI, Meta, and Google have agreed to label AI-generated images produced by their models.

There isn’t a foolproof way to detect synthetic content, however, as the labels can be removed. ®



Source link

Leave a Reply

Your email address will not be published. Required fields are marked *