Donald Trump shares fake AI images online

DW (English)

Ahead of the US presidential elections in November, Republican candidate Donald Trump is sharing AI-generated content on social media. Is that illegal? Or unethical?

Transcript

00:00Is US singer Taylor Swift endorsing Donald Trump?

00:03Is US presidential candidate Kamala Harris speaking at a communist event?

00:08Well, both pictures were created with AI.

00:11Both pictures convey a message that is untrue.

00:14And both were shared by Donald Trump, the other US presidential candidate.

00:18The Taylor endorsement one on his account on Truth, the Harris picture on X, formerly

00:24known as Twitter.

00:25Is that legal?

00:26Ethically acceptable?

00:28And is using AI to create misleading pictures the new way to campaign?

00:33US Republican candidate Donald Trump is not the first politician to paint a negative picture

00:38of an opponent with the help of AI.

00:40With this AI-generated post on X, Argentinian president Javier Millet famed his political

00:46adversary Sergio Massa as a communist in 2023.

00:50Massa's supporters, on the other hand, didn't shy away from using AI either, displaying

00:55him in heroic poses.

00:59As election campaigns now largely take place on social media, posts like these are not

01:03being controlled or checked by a publisher.

01:06This would be the case with a newspaper.

01:09In the US, Kamala Harris' campaign is trying to reach a young audience with meme-worthy

01:14content.

01:16But so far they haven't shared any AI content comparable to Trump's posts.

01:20They have been falsely accused of doing so, though.

01:24On Truth Social, Trump shared a post that claimed Harris had faked the size of a crowd

01:29at one of her rallies.

01:31Truth is, Trump himself seems to be the one using the technology much more.

01:36He shared AI images of a fake Swifties for Trump campaign, for example.

01:41This was to make fun of the actual Swifties for Kamala initiative, which has Taylor Swift

01:46fans campaigning on X for the presidential candidate.

01:50Apparently a very small label saying satire indicated that the pictures weren't real.

01:55AI seems to be omnipresent in many political campaigns.

01:58But is this legal?

02:01Well, there are no international rules governing the production and distribution of AI content

02:07and deepfakes.

02:08Platforms like Facebook and Instagram are developing technology to automatically label

02:12AI content, but their efforts haven't been very effective so far.

02:17This leaves the issue to local guidelines.

02:21In the US, only about 20 states have enacted regulations on the use of AI-generated false

02:26images in elections.

02:28However, legal experts doubt that these rules would even apply to the Trump posts mentioned

02:33before.

02:34Typically, these laws only ban depictions of someone doing or saying something in a

02:40convincing way.

02:42So what about the AI images of Swift and Harris?

02:46Well, lawmakers would likely argue that they don't look realistic enough to be covered

02:50by existing regulations.

02:53The Harris image is obviously exaggerated.

02:56And the Swift image doesn't directly depict an opposing candidate.

03:00Moreover, to take legal action against political AI posts in the US, it would need to be proven

03:06in court that they caused harm or injury to voters.

03:11Taylor Swift herself might have a claim, though, because her likeness was used in a

03:16post that falsely issues an endorsement.

03:19However, it remains to be seen if the singer will take legal steps.

03:23There is one application of AI during elections that is explicitly forbidden in the US.

03:29So-called robo-calls.

03:31In these, potential voters are contacted by phone or video call.

03:35Then AI-fabricated audio or video of a politician is used to ask for their support.

03:41But again, this is only illegal in the US.

03:45During India's election period this year, robo-calls were used on a massive scale, with

03:50many people potentially unaware of interacting with an AI clone.

03:55What does that mean for the future?

03:57You could argue that drastic tactics have always been used in election campaigns, especially

04:03when it comes to portraying political opponents in a negative light.

04:07But AI tools can produce tons of images much faster than human illustrators.

04:12And these images can easily be used strategically to create heated discussions online and distract

04:18from important political issues.

04:20What's more, AI generates imagery that looks deceivingly real.

04:25Which means generally we must doubt everything we see online first if we don't want to

04:30fall for fake news.

04:32I think it's a huge challenge for societies when people can't trust what they see and

04:38hear anymore.

04:40What's your opinion on this?

04:41Do we need strict rules for the use of generative AI, especially in a political context, or

04:47would that restrict freedom of expression?

04:50Let us know.

Category

Transcript

Recommended