Experts Warn of Election Interference From AI-Derived Disinformation

AI’s main impact in elections is through cognitive warfare, a battle for reshaping “hearts and minds.”

A photo illustration shows a user on an AI video generator website on a laptop computer in Taipei, Taiwan on Dec. 22, 2023. I-Hwa Cheng/AFP via Getty Images

Global experts are worried AI-derived disinformation will interfere with elections. This year, there will be approximately 70 general elections in more than 50 countries and regions worldwide.

The World Economic Forum (WEF) published an article at the end of January titled “The Big Election Year: how to Stop AI Undermining the Vote in 2024.” It raised concerns that the proliferation of AI-derived misinformation and disinformation would affect 4.2 billion people in nearly half of the planet, including the United States, India, the United Kingdom, Mexico, and Indonesia, who are going to cast a vote this year. The article has since been removed from the WEF website.

In early January, WEF’s Global Risks Report 2024 singled out “misinformation and disinformation” as a top risk, saying it could seriously destabilize the legitimacy of newly elected governments, bringing risks of political unrest, violence, and terrorism.

Depicting AI’s harm, it indicates, “No longer requiring a niche skill set, easy-to-use interfaces to large-scale artificial intelligence (AI) models have already enabled an explosion in falsified information and so-called ‘synthetic’ content, from sophisticated voice cloning to counterfeit websites.”

Josep Borrell, vice-president of the EU’s Executive Committee and High Representative for Foreign Affairs and Security Policy, stressed on Jan. 24 at a press briefing in Brussels that robust cooperation is needed to “fight FIMI (Foreign Information Manipulation and Interference) on a global scale,” given 2024 is a critical year for elections.

He warned, “Unchecked, malicious content spreads like a cancer and puts the health of our democracies at risk.”

CCP’s AI-Powered Cognitive Warfare

The Chinese Communist Party (CCP) interfered in Taiwan’s election in January. The CCP has long claimed sovereignty over the island and sought reunification by force.

Confetti flies over the stage and crowd as Taiwan's president-elect Lai Ching-te (Center L), and his running mate Hsiao Bi-khim (Center R) speak to supporters at a rally at the party's headquarters in Taipei, Taiwan on Jan. 13, 2024. Annabelle Chih/Getty Images

Despite Taiwan electing the anti-communist Democratic Progressive Party (DPP) faction of Lai Ching-Te, disinformation such as voting fraud orchestrated by the CCP affected the public’s trust in Taiwan’s democracy and newly elected government. The CCP AI-generated fake videos, pictures, and texts to achieve cognitive warfare, amplifying and spreading misinformation across online platforms.

On Jan. 10, Taiwanese national security sources revealed that AI-generated fake videos and audio had appeared on the internet through hundreds of fake accounts, featuring a virtual anchor reading from an e-book called “The Secret History of Tsai Ing-wen,” which is dedicated to spreading rumors about the Democratic Progressive Party (DPP) and then Taiwanese president Tsai Ing-wen.

According to a report published by Taiwan’s Information Environment Research Center on Jan. 23, a substantial number of high-traffic internet celebrities and private bloggers have circulated rumors across online platforms including Facebook, YouTube, and TikTok, alleging that DPP won the election through “voting fraud.”

The report mentioned that TikTok, commonly used by young people in Taiwan, was the hardest hit by fake messages. TikTok’s parent company in China, ByteDance, has been repeatedly linked to the CCP.

Ethan Tu, Taiwan AI Labs founder, told The Epoch Times the CCP used the internet to disseminate manipulative statements about the Taiwan Strait crisis, the Cross-Straits Economic Co-operation Framework Agreement (ECFA), and other issues in the CCP’s interest, “with the main purpose of giving the impression that the CCP is a peaceful representative while attempting to convince people to believe that the U.S. is the troublemaker that brought about the [potential] war.”

Mr. Tu also noted some CCP-controlled accounts “have clear schedules and starting and finishing times. They post repeatedly on the same issue simultaneously to increase netizen reactions, and they coordinate to suppress pro-Hong Kong, anti-communist comments, or other online comments against it.”

In disrupting Taiwan’s political situation and expanding its influence, the CCP applied AI-driven cognitive warfare in a bid to create public distrust, social division, and conflict escalation.

2024 Election Year 

In February, Indonesia, home to nearly 280 million, will elect a new president, and Pakistan will also hold parliamentary elections. In March, Russia will kick off presidential elections. In April and May, India, with a population of more than 1 billion, will have parliamentary elections. In June, there will be presidential elections in Mexico and parliamentary elections in the European Union, which represents a population of 500 million.

The most significant will be the U.S. presidential election on Nov. 5, given the United States’ leading position in global politics, economy and trade, the military, and its diplomatic policy with totalitarian states such as China, North Korea, Iran, and Russia.

Three competitive candidates, former President Donald Trump, current President Joe Biden, and former Democrat Robert F. Kennedy Jr., are moving forward with their campaigns.

In New Hampshire, a robocall of an AI-generated voice resembling President Biden urged supporters of the Democratic Party to vote for the Republican presidential candidate in the primary. In addition, some AI-produced videos and pictures about former President Trump have circulated online.

In response, New Hampshire Representative Angela Brennan warned, “You know, it’s not like maybe it used to be where it costs thousands of dollars to create a deep fake. We’re talking about something that someone can do on their phone in five minutes’ time. It’s a really big problem and it’s something we all need to be paying attention to.”

Americans receive more than 78 billion AI calls and more than 200 billion AI auto-dialed calls and automated texts each year, according to a phone scam report published by Robokiller, a mobile phone software that protects against fake calls and texts.

“Both big U.S. tech companies and security departments should take steps to stop AI meddling in the elections, especially interference by the CCP through various means,” Japanese electrical engineer Li Jixin told The Epoch Times.

Google, Meta, and other tech giants have announced measures to curb AI interference in elections in 2023, including labeling policies for AI to make it easier for people to identify it.

More countries hope to cope with the problems brought by AI by accelerating legislation.

OpenAI’s Solution

OpenAI CEO Sam Altman speaks during a keynote address announcing ChatGPT integration for Bing at Microsoft in Redmond, Washington, on Feb. 7, 2023. Jason Redmond/AFP via Getty Images

Sam Altman, CEO of OpenAI, a U.S.-based AI research organization, raised concerns about AI influencing U.S. and world elections late last year. He said, “What if an AI reads everything you have ever written online—every tweet, every article, every everything? And right at the exact moment sends you one message, customized just for you, that really changes how you think about the world.”

“That’s a new kind of interference that just wasn’t possible before AI,” he said.

On Jan. 15, OpenAI outlined how it plans to combat misinformation in the 2024 elections. Its plan is designed to prevent people with ulterior motives and hackers from using AI to generate incorrect and lifelike text and images, in order to mislead voters and influence the outcome of the election.

The new strategy designed “safeguards” against the image-generating AI DALL-E, rejecting requests to generate images of real people (including candidates) and prohibiting people from using its technology to create chatbots to distort the way voting works, or for political campaigning or lobbying purposes.

In addition, OpenAI will adopt digital authentication from the Content Source and Authenticity Alliance (C2PA), a cryptographic technology that encodes content sources with details to improve the identification and tracking of digital sources. C2PA members include Intel, Sony, Adobe, and Canon.

OpenAI also adds “digital watermarks” to AI-generated images to make them easier to recognize.

However, “There are still some immaturities in OpenAI’s proposed plans,” Kiyaohar Jin, a computer engineer in Japan, told The Epoch Times, “it remains to be seen whether these practices will actually catch AI abusers.”