China News Service, February 21 (Chen Caixia) "What a load of nonsense! The important thing is that you save your vote for the November election. This week's vote will only allow Republicans to re-elect Donald Trump... "Recently, before the primary election in New Hampshire, many Americans received a phone message from "President Biden", and this call was considered a "fake call" generated by artificial intelligence (AI).
With the advent of the era of generative artificial intelligence (GAI), "seeing is believing" and "pictures and truth" are gradually being subverted. At the beginning of 2024, the American company OpenAI released the first Vincent video model "Sora". While exciting the technology circle, it also made people increasingly worried about the proliferation of fake videos and pictures generated by AI. The emergence of a large number of "deep fake" videos may have a negative impact on political and security fields such as elections. Countries around the world still need to cooperate to deal with this challenge.
Image source: Screenshot of OpenAI website
It’s hard to tell whether “AI face-changing” is real or fake
Politicians from many countries "laying guns"
In recent years, AI technology has developed rapidly, and tools for creating pictures and videos such as Pika and Sora have emerged one after another. The "AI face-changing" is also processed through deep synthesis algorithm, providing support for "secondary creation". Deep forgery technology seems to be getting more and more out of the ordinary, and even politicians from many countries are taking risks...
On March 18, 2023, former US President Trump posted on social media that he expected to be arrested on March 21, 2023. As soon as this news came out, the "photo" of "Trump's arrest" quickly spread on some social platforms. These "photos" include "Trump being rounded up by the police", "Trump wearing prison uniform", and even "Trump escaping from prison", which are as exciting as Hollywood blockbusters.
Japanese Prime Minister Fumio Kishida was also criticized for the "AI face-changing" technology. In November 2023, a 25-year-old man in Japan used AI technology to create a fake video of "Kishida Fumio's press conference".
"Kishida" in the video is wearing a suit, with a nervous expression and a solemn tone. The video also features the Nippon TV station's logo and the words "Breaking News." This fake video has been viewed more than 2.3 million times in just one day since it went online.
Screenshot of the fake video of "Modi Dancing" Image source: Screenshot of social platform X
Indian Prime Minister Modi has also been involved in the "deep forgery" controversy. In November 2023, a video of "Modi dancing" went viral on the Internet. In the video, "Modi" and other dancers, wearing traditional costumes, formed a circle and happily danced the Garba dance, a festive dance in the Indian state of Gujarat.
Modi personally refuted the rumors, saying that the video was synthesized by AI. He warned, "People will mistakenly believe that what they see is real, which will cause a huge crisis... In the era of artificial intelligence, it is important to use technology responsibly."
Singapore Prime Minister Lee Hsien Loong also posted on social media at the end of 2023 that some deep fake videos using his audio had appeared recently to promote cryptocurrency scams. In the fake video circulating, Lee Hsien Loong mentioned investment opportunities approved by the Singapore government and promoted a so-called "revolutionary investment platform" designed by Tesla CEO Musk. Lee Hsien Loong called on the public to be more vigilant and not to be deceived.
Is the moment of political “deep hypocrisy” coming?
AI security becomes the focus
In addition to satisfying the novelty and social needs of online audiences, AI technology has gradually emerged in elections recently. Its integration with politics has brought about both progress and "traps."
Vallas denounced false videos posted online about him. Image source: Screenshot of CBS report
In February 2023, on the eve of the Chicago mayoral election in the United States, a video of "moderate Democratic candidate Paul Vallas lashing out at Chicago" quickly circulated on the Internet. In the video, Valas said that due to the lack of law and order, Chicago has become a "lawless" vacuum zone, and "the police can shoot suspects casually, but people turn a blind eye to these situations."
However, this video was proven to be artificially synthesized. Although the social platform finally deleted the video, the negative impact it had could not be eliminated, and Valas also lost to his opponent Brandon Johnson in the mayoral election.
"USA Today" quoted Darrell West, a senior fellow at the Brookings Institution's Center for Technology Innovation, as saying, "In the 2024 U.S. election, there will be a 'tsunami' of false information generated by artificial intelligence (AI). We We’ve seen this happen and it’s going to get worse.”
Robert Weisman, head of the U.S. government watchdog organization Defending the Public Interest, also said that the political "deep fake" moment has arrived and policymakers must take protective measures as soon as possible, otherwise they will face "election chaos."
Concerns that AI may be used to mislead voters and maliciously distort elections have risen sharply in the "global super election year" of 2024.
On February 16, many of the world's leading technology companies, including OpenAI, Amazon, Google, Microsoft, TikTok, Facebook's parent company "Yuan", and social platform The agreement was signed at the Security Council, committing to reduce the risk of using artificial intelligence to generate and distribute deceptive content to interfere with elections in 2024, a year when many countries will hold important elections.
"The government needs to play an important role in the field of artificial intelligence and ensure that safety barriers and guardrails are set up to let industry insiders know what can and cannot be done." said Shapiro, CEO of the Consumer Technology Association. (over)