Deepfakes: The 2024 US election is colliding with the rise of AI

“I actually like Ron DeSantis a lot,” Hillary Clinton reveals in a startling online video supporting the Republican candidate. “He’s exactly the kind of person this country needs, and I mean it.”

United States President Joe Biden shows himself as he is with a cruel attack on a transgender person. “You will never be a real woman,” grumbles the president.

In the 2024 United States presidential race, reality is at stake.

Clinton and Biden’s deepfake videos – realistic but fabricated, created by artificial intelligence algorithms trained from online material – are among thousands appearing on social media, blurring reality and fiction in the polarized world of American politics.

While these types of synthetic media have been around for several years, they have gained momentum over the past year thanks to a slew of new “generative AI” tools, like Midjourney, that make it easy and inexpensive to create convincing counterfeits, according to interviews. from Reuters with about two dozen specialists in areas such as artificial intelligence, online disinformation and political activism.

“It is going to be very difficult for voters to distinguish what is true from what is false. And one could imagine how Trump supporters or Biden supporters could use this technology to make the adversary look bad,” said Darrell West, senior fellow at the Brookings Institution’s Center for Technology Innovation.

“There could be things that come out just before the election that no one has the ability to deny.”

Tools that can generate deepfakes or deepfakes are being launched with few or no imperfect barriers to prevent harmful misinformation as the tech industry engages in an AI arms race, said Aza Raskin, co-founder of the Center. for Human Technology, a non-profit organization that studies the impact of technology on society.

Former President Donald Trump, who will compete with DeSantis and others for the Republican nomination against Biden, himself shared a doctored video of CNN host Anderson Cooper earlier this month on his social media platform TruthSocial.

“It was President Donald J. Trump kicking our ass here on CNN’s live presidential forum,” Cooper said on the recording, though the words don’t match his lip movement.

CNN said the video was fake. A representative for Trump did not respond to a request for comment on the clip, which was still on his son Donald Jr’s Twitter page this week.

While major social media platforms like Facebook, Twitter, and YouTube have made efforts to ban and remove deepfakes, their effectiveness in monitoring content varies.

Three times as many deepfake videos of all genres and eight times as many voice manipulations have been uploaded this year than at the same time in 2022, according to DeepMedia, a company working on synthetic media detection tools.

In total, DeepMedia estimates that some 500,000 video and voice impersonations will be shared globally on social media by 2023. Until late last year, cloning a voice cost $10,000 in training on servers and AI, but now companies are offering it for just a few dollars, he says.

According to those interviewed, no one knows for sure where generative AI is heading or how to effectively protect against its power of mass misinformation.

Industry leader OpenAI, which has changed the game in recent months with the release of ChatGPT and the updated GPT-4 model, is grappling with this problem. Its chief executive, Sam Altman, told the US Congress this month that election integrity was a “significant area of ​​concern” and called for swift regulation of the sector.

Unlike other smaller companies, OpenAI has taken steps to restrict the use of its products in politics, according to a Reuters analysis of the terms of service of half a dozen leading companies that offer software services. generative artificial intelligence. However, the barriers have flaws.

For example, OpenAI says it forbids its DALL-E image generator from creating public figures. When Reuters attempted to create images of Trump and Biden, the request was blocked and a message appeared saying it “may not follow our content policy.”

However, Reuters was able to create images of at least a dozen other US politicians, including former Vice President Mike Pence, who is also eyeing a run for the White House in 2024.

OpenAI also restricts any “large-scale” use of its products for political purposes. This prohibits, for example, the use of its AI to send personalized mass emails to voters.

The Microsoft-backed company explained its policies to Reuters in an interview but did not respond to other requests for comment on flaws in the application of its rules, such as blocking the creation of images of politicians.

Many small businesses do not have explicit restrictions on political content.

Launched last year, Midjourney is the leading company in AI-generated imagery, with 16 million users on its official Discord server. According to four AI researchers and creators interviewed.

Midjourney did not respond to a request for comment for this article. During a Discord chat last week, CEO David Holz said the company would likely make changes ahead of the election to combat misinformation.

Midjourney wants to cooperate on an industry solution to enable traceability of AI-generated images with a digital equivalent of the watermark and would consider blocking images of political candidates, Holz added.

As the industry strives to prevent abuse, some politicians are trying to harness the power of generative AI to improve their campaigns.

So far, the only major AI-generated political ad in the United States was one published by the Republican National Committee in late April. The 30-second ad, which the RNC revealed was entirely AI-generated, used fake footage to suggest a cataclysmic scenario if Biden is re-elected, with China invading Taiwan and San Francisco taken over by the crime.

The Republican National Committee did not respond to requests for comment on the announcement or its widespread use of AI. The Democratic National Committee declined to comment on its use of the technology.

Reuters polled all Republican presidential campaigns on their use of AI. Most didn’t respond, though Nikki Haley’s team said they weren’t using the technology and candidate Perry Johnson’s campaign said they were using AI for “copy generation.” and iteration”, without giving more details.

The potential of generative AI to compose emails, messages and campaign ads is compelling to some campaigners who believe this low-cost technology could level the playing field in elections.

Political consultants are also trying to harness AI, further blurring the line between the real and the unreal.

Numinar Analytics, a political data company that focuses on Republican customers, has begun experimenting with AI content generation for audio and images, as well as voice generation to create personalized messages with voice over. a candidate, founder Will Long said in an interview.

Democratic polling and strategy group Honan Strategy Group is trying to develop an AI-powered polling bot. Its chief executive, Bradley Honan, says he hopes to have a female robot in time for the 2023 municipal elections, citing studies showing that both men and women are more likely to speak to a female interviewer.